From Application-Driven Growth to Paradigm Shift: Scientific Evolution and Core Bottleneck Analysis in the Field of UAV Remote Sensing

Abstract

Unmanned Aerial Vehicle Remote Sensing (UAV-RS) has emerged as a transformative technology in high-resolution Earth observation, with widespread applications in precision agriculture, ecological monitoring, and disaster response. However, a systematic understanding of its scientific evolution and structural bottlenecks remains lacking. This study collected 4985 peer-reviewed articles from the Web of Science Core Collection and conducted a comprehensive scientometric analysis using CiteSpace v.6.2.R4, Origin 2022, and Excel. We examined publication trends, country/institutional collaboration networks, keyword co-occurrence clusters, and emerging research fronts. Results reveal an exponential growth in UAV-RS research since 2015, dominated by application-driven studies. Hotspots include vegetation indices, structure from motion modeling, and deep learning integration. However, foundational challenges—such as platform endurance, sensor coordination, and data standardization—remain underexplored. The global collaboration network exhibits a “strong hubs, weak bridges” pattern, limiting transnational knowledge integration. This review highlights the imbalance between surface-level innovation and deep technological maturity and calls for a paradigm shift from fragmented application responses to integrated systems development. Our findings provide strategic insights for researchers, policymakers, and funding agencies to guide the next stage of UAV-RS evolution.

1. Introduction

Unmanned Aerial Vehicle Remote Sensing (UAV-RS) has evolved from a niche technology into a mainstream scientific tool for high-resolution Earth observation using UAV platforms [1,2]. Compared to traditional satellite and manned aerial remote sensing, UAV-RS offers significant advantages in spatiotemporal resolution, acquisition efficiency, and cost-effectiveness, making it particularly suitable for geographic scenarios requiring high-frequency observation, fine-scale monitoring, and rapid response.

In recent years, with the advancement of key technologies such as flight control, navigation and positioning, and remote sensing sensors, the application domains of UAV-RS have continuously expanded. UAV-RS now shows tremendous potential in areas such as precision agriculture [3,4], ecological and environmental monitoring [5], disaster emergency response [6], urban planning [7], marine surveillance [8], and infrastructure management [9]. For example, in agriculture, UAV-RS can be used for crop health monitoring [10], soil nutrient assessment [11], and pest and disease warning [12], facilitating the digital and precision transformation of agriculture. In the ecological domain, UAV-RS enables dynamic monitoring of forests [13], wetlands [14], and water bodies [15], providing vital technical support for resource conservation and environmental governance. In disaster management, UAV-RS allows rapid acquisition of high-resolution imagery to support emergency assessment and decision-making [16]. In urban management, it can be applied to land use monitoring [17], urban expansion analysis [18], and traffic control [19].

Although multiple review articles on UAV-RS have emerged in recent years [1,20,21,22,23,24], most focus on specific technologies or application fields [25], and there is still a lack of systematic review and quantitative analysis of the overall development trajectory of UAV-RS. For instance, many studies emphasize sector-specific applications such as agriculture [26,27], forestry [28,29], fisheries [8], archaeology [30], natural disasters [31], and ecological monitoring—including grasslands [32], wildlife [33], and biodiversity [34]—yet cross-disciplinary integrations are rarely discussed. Moreover, most reviews tend to emphasize technical aspects like sensor design and image processing [35], with limited attention to global research landscapes, collaborative networks, and technological convergence trends in UAV-RS.

Despite the exponential growth of the UAV-RS literature, a critical tension has become increasingly evident: has the rapid expansion on the application side masked stagnation on the technological front? In other words, while a large body of research focuses on how to efficiently apply UAVs to address practical problems in agriculture, ecology, and related fields, have the foundational technologies that support these applications—such as platform endurance, sensor coordination, data standardization, and global collaboration mechanisms—quietly become bottlenecks limiting the field’s long-term development?

To address this, the present study proposes three interrelated core questions that serve as the analytical framework for this paper:

(1)

Does quantitative evidence indicate that UAV-RS research has become predominantly driven by application demands, at the expense of investment in foundational research?

(2)

Within the global landscape of knowledge production, is there evidence of “technological silos” that hinder deep collaboration across institutions and across national or regional boundaries?

(3)

Behind the current enthusiasm for multi-source data fusion, is the lack of unified data standards and interoperability frameworks posing hidden risks to data integration?

Answering these questions not only helps to clarify the internal tension between the rapid development of application-level UAV-RS and the relative stagnation of its technical foundations, but also provides forward-looking strategic insights for researchers, funding agencies, and policy makers.

2. Materials and Methods

2.1. Literature Selection

To ensure the representativeness and feasibility of the research findings, this study selected the Web of Science (WOS) Core Collection as the source of the literature data. Based on existing research experience and standardized terminology, the search strategy was formulated as follows: TS = (“unmanned aerial vehicle” OR “unmanned aircraft vehicle” OR UAV OR drone* OR “unmanned aircraft*” OR “unmanned helicopter*” OR “unmanned aerial system*” OR “unmanned aircraft system*” OR “pilotless aircraft*” OR “pilotless helicopter*” OR UAS) AND TS = (“remote sensing”). This search query incorporates commonly used UAV-related terms and the theme of remote sensing to ensure comprehensive coverage of the relevant literature. The SCI-EXPANDED database focuses on the natural sciences, while the SSCI focuses on the social sciences. The citation indices selected include the Science Citation Index Expanded (SCI-EXPANDED, 1900–present) and the Social Sciences Citation Index (SSCI, 1956–present). The search was conducted up to 31 December 2024, yielding a total of 7118 relevant publications. The document types were limited to peer-reviewed journal articles and reviews (Article, Review), excluding editorial materials, meeting abstracts, book chapters, corrections, and news items. Through a combination of manual screening and automated filtering, irrelevant and duplicate records were removed. The final dataset included 4985 UAV Remote Sensing (UAV-RS) publications, spanning multiple disciplinary domains and research directions, thus ensuring both representativeness and temporal relevance of the data for scientometric analysis (Figure 1). Our search strategy did not manually restrict the starting year. The Science Citation Index Expanded (SCI-EXPANDED) covers publications dating back to 1900, but this does not represent the actual starting point of our analysis. Instead, our study was based on the default coverage range of the database to ensure that no potentially relevant early studies were overlooked.

2.2. Research Methods

Scientometrics is a discipline closely related to scientific exploration and evaluation, primarily focusing on the quantitative and visual analysis of research quality, scientific trends, and the influence of research domains [36]. The use of scientific knowledge mapping methods facilitates the generation of new knowledge and discoveries [37]. In this study, CiteSpace 6.1.3 was employed [38], a widely used tool for bibliometric analysis in specific fields. It enables the exploration of research hotspots, emerging trends, key authors, and institutions through visual knowledge maps [39], providing valuable references for scientific decision-making and resource allocation. The software has gained extensive adoption among researchers worldwide [40,41,42]. In this analysis, CiteSpace, Origin, and Microsoft Excel were used to conduct statistical analyses and generate visualizations. First, annual publication trends and contributions from different countries, institutions, and journals were quantified to outline the overall research landscape. Next, a keyword co-occurrence analysis was performed to construct a thematic research network, where clustering algorithms were used to identify the major fields of UAV-RS research. Finally, a temporal trend analysis was conducted by detecting burst keywords, aiming to identify emerging topics and frontier trends that have gained significant attention over time. This quantitative analysis framework provides a solid and objective foundation for interpreting the evolution and current state of the UAV-RS research field. By integrating traditional scientometric methods into a diagnostic analytical framework to address deep structural issues in the UAV-RS field that have not yet been systematically quantified, this study provides a solid and objective foundation for interpreting the evolution and current state of the field.

The bibliometric analysis was conducted using CiteSpace 6.1. The main parameter settings and results were as follows: Timespan: 1993–2024 (Slice Length = 1); Selection Criteria: g-index (k = 25), LRF = 3.0, L/N = 10, LBY = 5, e = 1.0; Network: N = 542, E = 2674 (Density = 0.0182); Largest 30 CCs: 515 (95%); Nodes Labeled: 1.0%; Pruning: None; Modularity Q = 0.4886; Weighted Mean Silhouette S = 0.8007; Harmonic Mean (Q, S) = 0.6068.

3. Results

3.1. Temporal Evolution

The annual publication volume of literature serves as a key indicator for evaluating the level of activity and developmental stage of a research field. UAV-RS technology has undergone a long evolutionary process, transitioning from exclusive military use to widespread civilian adoption. Its scientific evolution vividly reflects the technological advancements and integration across multiple domains, including aerospace, sensor technology, information processing, and artificial intelligence. This evolution is not only evident in the performance enhancement of UAV platforms themselves, but also in innovations in sensor technology, the increasing intelligence of data processing and analysis methods, and the continuous expansion and deepening of application areas. In its early stages, UAV-RS was primarily used for military reconnaissance, leveraging its advantages such as pilotless operation and the ability to access high-risk areas, thereby playing a vital role in intelligence gathering. Entering the 21st century, with the maturation of technology and a significant reduction in costs, UAV-RS began to penetrate civilian sectors, demonstrating enormous potential in industries such as agriculture, forestry, environmental monitoring, disaster response, and surveying and mapping. In particular, the emergence of lightweight and compact UAVs has significantly lowered the threshold for UAV-RS, enabling more enterprises and individuals to use this technology for data collection and analysis. For instance, the release of the DJI Phantom 1 in 2012 by China’s DJI company marked a milestone. With its user-friendly design and relatively low cost, it quickly captured the consumer market and facilitated the use of UAV-RS in crowdsourced geographic data acquisition [43]. This shift signifies the transition of UAV-RS from a specialized, high-cost military application to a more democratized, low-cost civilian service. Its scientific evolution path thus exhibits distinctive characteristics of multi-technology integration and cross-sectoral penetration. The annual publication trend is shown in Figure 2.

(1) Initial Stage (Before 2000): During this period, the number of UAV-RS-related publications was extremely limited, with annual outputs generally in the single digits. Research was primarily driven by early explorations in military and aerospace engineering contexts, and both the application scenarios and technical capabilities of UAVs remained in an experimental phase.

(2) Slow Growth Stage (2000–2014): With the proliferation of GPS navigation, inertial measurement units (IMUs), and lightweight flight platforms, UAVs gradually began to enter scientific and commercial domains. Around 2010, the annual number of publications surpassed 20, reaching 84 by 2014. Representative studies during this period were largely focused on agricultural remote sensing, disaster monitoring, and photogrammetry applications. Although recognition of UAV-RS steadily increased, progress remained relatively slow due to limitations in cost, platform stability, and data processing capacity.

(3) Rapid Development Stage (2015–2019): After 2015, civilian UAV technologies advanced rapidly. In particular, commercial aerial platforms—led by companies like DJI in China—significantly lowered technical barriers and improved operability, offering low-cost and high-quality data acquisition solutions for researchers. Annual publication volume exceeded 100 papers for the first time in 2016 and surged to 470 by 2019, with an average annual growth rate of over 30%. During this phase, UAV-RS technologies achieved widespread deployment in agriculture, forestry, ecology, hydrology, and more. Research themes shifted from “whether UAVs can be used” to “how to use them effectively.”

(4) High-Activity Stage (2020–Present): Since 2020, UAV-RS has maintained a high-growth momentum, with annual publications consistently exceeding 600. Research hotspots have further expanded into areas such as “multi-sensor integration,” “intelligent recognition,” “3D reconstruction,” and “deep learning,” reflecting increasingly frequent interdisciplinary collaboration. Meanwhile, global cooperation networks have been gradually expanding, forming a multi-level advancement pattern that spans theory, technology, and application. This exponential growth curve underscores the profound impact of UAV technology on scientific research.

3.2. Analysis of the Global Scientific Collaboration Network: Formation of a Dual-Core Structure

3.2.1. Global Distribution of National/Regional Research Capacities

Based on the 4985 articles retrieved and screened from the Web of Science database, this study conducted a statistical analysis of publication volume by country/region. These publication counts were assigned to corresponding geographic vector in ArcGIS 10.2. For visual clarity and representativeness, only countries/regions with more than 10 publications were included in the map. An analysis of national publication statistics and collaboration networks in the field of UAV-RS reveals a marked geographic concentration and high degree of cooperative activity (Figure 3). Researchers from over 40 countries and regions have contributed to this field; however, the top 15 countries account for nearly 85% of the total global publications. This indicates a highly concentrated academic output, suggesting the presence of significant entry barriers or a strong concentration of research resources in this domain.

China ranks first, with 1657 publications, accounting for 33.2% of the total (

Table 1. The top 15 countries/regions in terms of publication volume and their main research fields.

Rank	Countries/ Regions	Number of Publications	Typical Research Areas
1	China	1657	Agricultural applications [44,45], Environmental Monitoring [46], Urban Management [47], Disaster Monitoring [48], Cultural Heritage Protection [58], Water management [59], Wildlife Monitoring [60], Forest monitoring [61], Land Management [62]
2	USA	1094	Environmental Monitoring [49], Disaster Response [50], Crop Monitoring [51], Biodiversity monitoring [52], Cultural Heritage Protection [53], Urban Plan [54], Infrastructure Monitoring [55], Wildfire Damage Assessment [56], Forest monitoring [57]
3	United Kingdom	367	Agricultural applications [63], Land cover change [64], Coastline Change and Ocean Monitoring [65], Archaeological exploration [66], Urban management monitoring [67]
4	Italy	340	Agricultural applications [68], Environmental Monitoring and Disaster Response [69], Heritage Conservation and Archaeology [70], Urban infrastructure monitoring [71], Forest monitoring [72], Volcano monitoring [73]
5	Germany	308	Agricultural applications [74], Environmental Monitoring [75], Urban planning and infrastructure monitoring [76], Geological Exploration [77], Forest monitoring [78]
6	Australia	284	Environmental Monitoring and Ecological Protection [79], Forestry Ecosystem Research [80], Water Resources and Coastline Research [81], Agricultural applications [82], Natural disaster monitoring [83], Land Management [84]
7	Canada	209	Disaster Response [85], Agricultural applications [86], Urban infrastructure monitoring [87], Archaeological research [88], Wildlife Monitoring [89], Forest monitoring [90]
8	Brazil	177	Land use and land cover change [91], Tropical forest monitoring and ecosystem research [92], Forest fire monitoring [93], Agricultural applications [94]
9	France	163	Agricultural applications [95], Urban infrastructure monitoring [96], Archaeological research [97], Coastal environmental monitoring [98]
10	Netherlands	135	Agricultural applications [99], Environmental Monitoring and Biodiversity Conservation [100], Urban Management [101]
11	South Korea	128	Agricultural applications [102], Disaster Monitoring [103], Beach and water quality monitoring [104], Forest monitoring [105]
12	Japan	122	Agricultural applications [106], Environmental Monitoring [107], Urban Plan [108], Forest monitoring [109], Volcano monitoring [110]
13	India	107	Agricultural applications [111], Forest monitoring [112], Urban Management [113], Land use monitoring [114], Glacier Survey [115], Coastline monitoring [116]
14	Portugal	89	Agricultural applications [117], Environmental Monitoring [118], Biodiversity Conservation [119], Urban Management [120], Land Use [121], Disaster Monitoring [122], Forest monitoring [123]
	Switzerland	89	Agricultural applications [124], Environmental Monitoring [125], Urban Plan [126], Snow monitoring [127]
15	Finland	86	Agricultural applications [128], Environmental Monitoring [129], Archeology [130], Forest monitoring [131]

). Its research spans various themes such as agricultural remote sensing [44,45], ecological monitoring [46], urban management [47], and disaster response [48]. The United States ranks second, with 1094 publications (21.9%), with a focus on ecological environments, disaster assessment, and urban infrastructure monitoring [49,50,51,52,53,54,55,56,57], reflecting a strong orientation toward applied technologies. Together, China and the U.S. contribute more than half of the global academic output in the UAV-RS field, forming a dominant first-tier group. China’s research agenda is distinctly “application-driven,” emphasizing precision agriculture, environmental monitoring, and infrastructure inspection, closely aligned with national strategic priorities. While the U.S. is also application-focused, it demonstrates a deep foundation in ecosystem monitoring, disaster evaluation, and fundamental remote sensing methodologies. European countries such as the UK, Italy, and Germany—traditionally strong in remote sensing—comprise the second tier, with research that emphasizes regional characteristics (e.g., archaeology, volcanic monitoring) and the refinement of specific technologies such as LiDAR.

3.2.2. Institutional Cooperation

Studies reveal that UAV-RS research is highly dependent on a tripartite collaboration mechanism involving universities, research institutes, and government agencies. Among the top 15 institutions by publication volume, the Chinese Academy of Sciences leads with 310 publications, underscoring its dominant role in fundamental theory development and system integration. Institutions such as the Ministry of Agriculture and Rural Affairs of China (148 papers), Chinese Academy of Agricultural Sciences (83 papers), and Nanjing Agricultural University (53 papers) are also highly productive, reflecting China’s strategic deployment in agricultural and ecological remote sensing. The U.S. Department of Agriculture (USDA) ranks fourth, with 120 papers, focusing primarily on precision agriculture and forestry management. The University of California system (96 papers) and the Texas A&M University system (54 papers) continue to contribute significantly in remote sensing method innovation and platform integration.

A noteworthy and concerning observation is the relatively weak direct collaboration between the two dominant research powers—China and the United States. Despite their enormous academic output, the intensity and quantity of direct collaborative links (represented by line thickness and connection count) are disproportionately low. This fragmented cooperation pattern may hinder the exchange of knowledge and data required to address global challenges such as climate change and food security, and may exacerbate barriers to technology transfer. In addition, European research institutions—including the Helmholtz Association in Germany, the French National Centre for Scientific Research (CNRS), and the Italian National Research Council (CNR)—have made significant academic contributions in LiDAR-based 3D modeling, cultural heritage preservation, and urban remote sensing.

Through an in-depth analysis of institutional collaboration and research domains in the field of UAV-RS (Figure 4,

Table 2. Top 15 research institutions by number of publications.

Rank	Institution	Countries/Regions	Number of Publications
1	Chinese Academy of Sciences	China	310
2	Ministry of Agriculture Rural Affairs	China	148
3	University of Chinese Academy of Sciences	China	140
4	United States Department of Agriculture	USA	120
5	Wuhan University	China	115
6	University of California System	USA	96
7	Chinese Academy of Agricultural Sciences	China	83
8	China Agricultural University,	China	76
9	Beijing Academy of Agriculture Forestry Sciences, Consejo Superior De Investigaciones Cientificas Helmholtz Association, Helmholtz Association	China, USA, Germany	73, 73, 73
10	Consiglio Nazionale Delle Ricerche	Italy	69
11	Institute of Geographic Sciences and Natural Resources Research, State University System of Florida	China, USA	67, 67
12	Centre National De La Recherche Scientifique	France	62
13	Beijing Normal University, Northwest A&F University	China	57, 57
14	Texas A&M University System, University of Florida	USA, China	54, 54
15	Nanjing Agricultural University	China	53

), several key patterns emerge. In China, the Chinese Academy of Sciences (CAS) and Wuhan University are the top two institutions in terms of publication volume. These institutions maintain close partnerships with various domestic organizations and have also established links with U.S. universities and institutions in several European countries. The CAS plays a central role in UAV-RS research, engaging in extensive collaboration both domestically and internationally. Its research domains span foundational theories of UAV-RS, UAV system design and applications, as well as remote sensing data processing and analysis.

In the United States, the United States Geological Survey (USGS) and the United States Department of Agriculture (USDA) exhibit strong collaborative ties, aligning with their research priorities in geological exploration and agricultural monitoring. These agencies have conducted extensive collaborative research on the application of UAV-RS technologies in geospatial surveys, land resource assessment, and crop monitoring. Additionally, the involvement of NASA and the U.S. Department of the Interior highlights the importance of UAV-RS in environmental protection, disaster prevention, and space exploration. These institutions jointly develop and apply UAV-RS technologies to enhance environmental monitoring and Earth observation. Compared with China, U.S. institutions show a more balanced spatial distribution and maintain closer inter-institutional linkages, fostering frequent collaborations across organizations.

European research institutions such as the Centre National de la Recherche Scientifique (CNRS) in France and Wageningen University and Research in the Netherlands also hold prominent positions in the global collaboration network, particularly in areas such as precision agriculture, soil property analysis, and geoscience.

It is also noteworthy that the collaboration network includes institutions focused on niche technical domains such as structure from motion (SfM), soil surface characteristics, and precision nitrogen management. These trends suggest that UAV-RS is increasingly intersecting with multiple disciplinary fields, resulting in a more expansive and integrated research collaboration ecosystem.

3.3. Journals and Authors

3.3.1. Journals

Academic journals serve as critical platforms for disseminating research findings, and the volume and quality of publications in these journals often reflect the research intensity and developmental trajectory of a given field (

Table 3. Top 10 periodicals publishing UAV-RS research.

Rank	Journal	Number of Publications
1	Remote Sensing	1068
2	Sensors	205
3	Drones	154
4	Computers and Electronics in Agriculture	118
5	International Journal of Remote Sensing	94
6	IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing	94
6	ISPRS Journal of Photogrammetry and Remote Sensing	94
7	Frontiers in Plant Science	88
8	International Journal of Applied Earth Observation and Geoinformation	86
8	Remote Sensing of Environment	86
9	Agronomy-Basel	83
10	IEEE Transactions on Geoscience and Remote Sensing	74

). Overall, research on UAV-RS exhibits clear trends toward diversification and specialization. The journal Remote Sensing leads by a wide margin, publishing 1068 articles and emerging as the primary outlet for UAV-RS research. Its prominence can be attributed to its broad scope and comprehensive coverage, spanning foundational theory, technical applications, case studies, and methodological innovation. This diversity makes it a key venue for scholars from various UAV-RS subfields to share and exchange ideas.

In addition to Remote Sensing, other journals such as Sensors, Drones, and Computers and Electronics in Agriculture have also published a substantial number of UAV-RS articles, each with a distinct thematic emphasis reflecting the growing demand and evolving trends of UAV-RS in different domains. Sensors ranks second with 205 articles, which is consistent with the journal’s focus on the development and application of sensor technologies within UAV-RS. As a core component of UAV-RS systems, advancements in sensor technology are vital for enhancing the accuracy and efficiency of data collection.

The journal Drones, with 154 articles, is dedicated specifically to UAV-related topics, including design, manufacturing, control, and diverse applications. Its publication volume underscores the research community’s strong interest in UAV platform development and operational innovation. Computers and Electronics in Agriculture ranks fourth with 118 publications, likely due to its focus on agricultural applications of UAV-RS. Agriculture is one of the most significant domains for UAV-RS, with research encompassing crop monitoring, pest and disease control, and land resource management.

Other notable journals include the International Journal of Remote Sensing and the IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, both of which contribute significantly to the field. These journals showcase the potential of UAV-RS in a wide range of applications such as environmental monitoring, urban planning, and disaster prevention, further emphasizing the interdisciplinary relevance and growing impact of UAV-RS research.

3.3.2. Authors

Overall, UAV-RS research has garnered widespread global attention from the scientific community. According to

Table 4. Publication numbers and influence of the top 10 authors in UAV-RS research.

Rank	Author	Number of Publications	Institution	Country
1	Haikuan Feng	43	Beijing Academy of Agriculture and Forestry Sciences	China
2	Guijun Yang	42	Beijing Academy of Agriculture and Forestry Sciences	China
3	Honkavaara E	36	National Land Survey of Finland	Finland
4	Yubin Lan	30	Shandong University of Science and Technology	China
5	Lucieer Arko	28	University of Tasmania	Australia
6	Yan Zhu	24	Nanjing Agricultural University	China
6	Weixing Cao	24	Nanjing Agricultural University	China
7	Wenting Han	21	Northwest A&F University	China
8	Matese Alessandro	20	University of Turin	Italy
8	Di Gennaro Salvatore Filippo	20	Consiglio Nazionale delle Ricerche	Italy
9	Hakala Teemu A	19	Finnish Geospatial Research Institute	Finland
10	Yongchao Tian	18	Henan Polytechnic University	China

, Chinese researchers have demonstrated outstanding research output in this field. Haikuan Feng from the Beijing Academy of Agriculture and Forestry Sciences leads with 43 published papers, closely followed by Guijun Yang with 42 publications. This indicates that the Beijing Academy of Agriculture and Forestry Sciences possesses strong capabilities and significant influence in UAV-RS, particularly in agricultural and forestry monitoring and land resource management.

Eija Honkavaara from the Finnish Geospatial Research Institute ranks third with 36 publications, reflecting her expertise in the application of UAV-RS in land surveying and geographic information systems (GIS). Other highly influential researchers in the field include Arko Lucieer from the University of Tasmania, Yan Zhu and Weixing Cao from Nanjing Agricultural University, and Yongchao Tian from Henan Polytechnic University in China. Their work spans various aspects of UAV-RS, including precision agriculture, environmental monitoring, forestry management, and remote sensing data analysis.

The affiliations and countries of these leading researchers highlight China’s robust research capacity and strong talent pool in UAV-RS. This aligns with the country’s strategic focus on agricultural modernization, environmental protection, and disaster prevention. Meanwhile, countries like Finland and Australia also maintain a notable presence in the field, reflecting their unique strengths and contributions to UAV-RS technology development.

From an annual perspective, research in the field of UAV-RS has shown a rapid growth trend. This can be attributed to the swift advancement of UAV technology, the continuous expansion of application domains, and the increasing global demand for precision agriculture, environmental protection, and disaster monitoring. Different researchers focus on distinct aspects within UAV-RS, and over time, their research interests and areas of expertise have also evolved (Figure 5). For example, Haikuan Feng and Guijun Yang—both prolific authors—have covered a wide range of UAV-RS topics in their work. Their research spans from soil salt content to vegetation index analysis, reflecting their in-depth exploration and contribution to agricultural remote sensing.

Researchers’ interests have gradually expanded from early studies in agricultural remote sensing to encompass a wide range of fields, including environmental monitoring, forest management, and disaster prevention. Over time, both UAV technology itself (drones) and its applications in remote sensing have shown a steady upward trend in research attention. Moreover, domain-specific applications of UAVs—such as volcanic monitoring, hydrological station observation, and precision forestry—have emerged as key research focuses. Alongside these developments, technological advancements related to UAVs—particularly in image processing and data analysis—have also become prominent research hotspots. For instance, the application of convolutional neural networks (CNNs) and biosensors has played a crucial role in improving image interpretation and data analytics. Additionally, research on image co-registration and image correlation is vital for enhancing the accuracy and reliability of UAV-RS data. Through the researchers’ quantitative analysis of keywords, this study reveals that beneath the surface-level prosperity of rapid growth in UAV-RS applications lies a weak foundation in basic research. The research agenda is largely driven by immediate application needs, while critical foundational technologies—such as sensors and data standards that are essential for the field’s long-term development—receive insufficient attention.

3.4. Analysis of Research Fields

3.4.1. Research Field and Content

As an interdisciplinary technology, Unmanned Aerial Vehicle Remote Sensing (UAV-RS) research outputs are widely distributed across multiple academic fields. Overall, UAV-RS studies are highly concentrated in application-oriented and technological disciplines, particularly in Remote Sensing, Environmental Sciences, and Imaging Science (

Table 5. Top 5 discipline categories of UAV-RS publications.

Rank	Subject Category	Number of Publications
1	Remote Sensing	2107
2	Environmental Sciences	1947
3	Imaging Science Photographic Technology	1707
4	Geosciences Multidisciplinary	1559
5	Engineering	883

).

Specifically: (1) Remote Sensing leads, with 2107 published papers, ranking first among all disciplines. This field encompasses Earth observation, data acquisition and processing, and information extraction. The development of UAV-RS has significantly advanced both the depth and breadth of research and applications in remote sensing. (2) Environmental Sciences follows closely, with 1947 papers. Researchers in this domain utilize UAV-RS for ecological conservation, pollution monitoring, and climate change impact assessment. The flexibility and high-resolution imaging capabilities of UAVs make them powerful tools in environmental monitoring and evaluation, underscoring their growing importance in this area. (3) Imaging Science and Photographic Technology ranks third, with 1707 papers. Research in this domain typically focuses on the design of UAV-mounted imaging systems, the development of image processing algorithms, and the enhancement of image quality. UAV-RS has not only driven advancements in imaging and photographic technologies, but has also provided high-quality data sources for other disciplines. (4) Geosciences Multidisciplinary is ranked fourth, with 1559 papers. UAV-RS technologies have found widespread applications in geological exploration, topographic mapping, and natural disaster assessment. The high-resolution 3D data provided by UAVs are crucial for understanding surface processes and features of the Earth.

Additionally, the Engineering field has contributed 883 papers, indicating UAV-RS applications in civil engineering, mechanical engineering, and related areas of engineering design and implementation.

3.4.2. Keyword Co-Occurrence and Clustering Analysis

Literature keywords serve as a rapid and accurate means of revealing research topics, technical methodologies, and application domains. To further explore the structural content and thematic evolution of UAV-RS research, this study employed CiteSpace to perform a co-occurrence clustering analysis of high-frequency keywords. The minimum occurrence threshold was set to 20, resulting in 10 thematic clusters. The modularity value of the clustering network was Q = 0.4394 (>0.3), and the silhouette score was S = 0.6018 (>0.5), both indicating that the clustering results are valid and structurally clear (Figure 6). The UAV-RS research was categorized into the following 10 co-clustering groups: #0-vegetation indices, #1-structure from motion, #2-object-based image analysis, #3-remote sensing, #4-deep learning, #5-hyperspectral imaging, #6-ECOSTRESS, #7-aerial photography, #8-soil surface, #9-land surface phenology. The analysis included a total of 742 keyword nodes and 6360 connecting edges. Due to its flexibility, cost-effectiveness, and high-resolution imaging capabilities, UAV-RS has emerged as a significant branch within the broader field of remote sensing. These clusters reflect the expanding application of UAV-RS across diverse research areas and its integration with advanced data processing technologies (e.g., deep learning), thereby enhancing the analytical and practical value of remote sensing data.

Each connecting line represents the co-occurrence of two keywords within the same article, highlighting the thematic relationships between them. This study selected the most frequently occurring keywords within each cluster (

Table 6. Clusters and high-frequency words of 10 UAV-RS-related research.

Serial Number	Cluster	Keyword (Frequency)	Representative Research Direction
#0	Vegetation Indices	Machine Learning, Unmanned Aerial Vehicle (UAV), Biomass, Index, Yield, Chlorophyll Content, Images, Prediction, Aboveground Biomass, Growth, Soil, Spectral Reflectance	Crop monitoring and health assessment. Core application: Agriculture and ecological monitoring are the dominant focuses.
#1	Structure From Motion	UAV, Photogrammetry, Lidar, Structure From Motion, Accuracy, Models, Segmentation, Height, Topography, Point Clouds, Performance, River, Lidar Data, Image Processing, UAV Imagery, Density, Reconstruction, Tree Species Classification, Terrestrial, Digital Photogrammetry, Infrared Thermography, Terrain	Dominant 3D Modeling Technology: The frequency of SfM significantly exceeds that of LiDAR, indicating that cost-effectiveness remains the primary factor in technology selection, rather than absolute accuracy.
#2	Object-Based Image Analysis	Random Forest, Management, Dynamics, Variability, Scale, Airborne Lidar, Water, Patterns, Biodiversity, Support Vector Machine, Landscape, Plants, Google Earth Engine	Image segmentation and object-based classification: Extended applications of traditional remote sensing image analysis methods.
#3	Remote Sensing	Remote Sensing, Unmanned Aerial Vehicle, Classification, Imagery, Precision Agriculture, Model, Leaf Area Index, Forest, Management, Crop	Fundamental remote sensing models and monitoring methods: A concentrated manifestation of application-driven research, with precision agriculture being the strongest demand driver.
#4	Deep Learning	Vegetation Index, Object Detection, Deep Learning, Feature Extraction, Chlorophyll Content, Computer Vision, Unmanned Aerial Vehicles, Fusion, Semantic Segmentation, Image Segmentation, Image Classification, Convolutional Neural Networks, W-Net, Spatial Resolution, Information, Land Cover, Time Series, Framework, Vegetation Mapping	Intelligent recognition and automated data processing: High-frequency terms are primarily established computer vision (CV) models, with a notable absence of new model terminology specifically adapted to remote sensing characteristics.
#5	Hyperspectral Imaging	Algorithm, Retrieval, Calibration, Hyperspectral Imaging, Surface, Calibration, MODIS, Land Surface Temperature, Landsat, Atmospheric Condition, Hyperspectral Remote Sensing, Water Quality, Coastal, Fluorescence	Emerging sensor technologies: Focused on algorithms and calibration, but not yet widely applied at scale.
#6	Eco Stress	Climate, Traits, Landsat8, Uncertainty, Vegetation Cover, Surface Temperature, Spatial Analysis	Ecological stress response analysis.
#7	Aerial Photography	Cover, Image Analysis, Aerial Photography, Fluxes, Satellite Images	Aerial image interpretation.
#8	Soil Surface	Forest, Features, Erosion, Image	Soil and surface structure monitoring.
#9	Land Surface Phenology	Climate-Change, Plant Phenology, Land Surface Phenology, Chlorophyll-Aa, Inland Waters	Phenological changes and climate response.

). By analyzing the characteristics of the 10 clusters and their high-frequency keywords, the results indicate that UAV-RS is increasingly integrated with advanced technologies such as machine learning, deep learning, and hyperspectral imaging. This integration not only enhances the efficiency and accuracy of data acquisition, but also significantly improves the capabilities of data processing and analysis. UAV-RS is thus actively driving the advancement of remote sensing science and playing an increasingly vital role in domains such as precision agriculture, environmental protection, and climate change research. Notably, clusters directly associated with practical applications—such as #0 (vegetation indices), #7 (aerial photography), and #9 (land surface phenology)—dominate the field. This suggests that current research efforts are primarily focused on utilizing existing UAV platforms and sensors to address specific application-driven problems. However, the current research agenda appears heavily skewed toward short-term application value, with insufficient attention paid to addressing core bottlenecks that constrain the field’s long-term development. These include issues such as UAV platform endurance, sensor performance, and data standardization—critical areas that require greater strategic focus and investment [132,133].

3.5. Evolution of Research Frontiers from Keyword Bursts

Analyzing the evolution of emerging keywords reveals clear trends in the development of UAV-RS (

Table 7. Top 40 emerging UAV-RS keywords.

Keywords	Year	Strength	Begin	End	1993–2024
Aircraft	2000	5.48	2000	2018	▂▂▂▂▂▂▂▃▃▃▃▃▃▃▃▃▃▃▃▃▃▃▃▃▃▃▂▂▂▂▂
Imagery	2001	18.4	2009	2018	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▃▃▃▃▂▂▂▂▂
Image processing	2005	4.82	2005	2017	▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▃▃▃▃▃▃▃▂▂▂▂▂▂
Aerial photography	2006	5.86	2006	2016	▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▃▃▃▃▃▂▂▂▂▂▂▂
Unmanned aerial vehicle	2006	7.57	2010	2015	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▂▂▂▂▂▂▂▂
Crop	2007	9.77	2007	2016	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▃▃▃▃▂▂▂▂▂▂▂
Features	2008	4.49	2015	2018	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▂▂▂▂▂
Canopy temperature	2009	3.51	2009	2017	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▃▃▃▂▂▂▂▂▂
Water stress	2009	7.46	2016	2018	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▂▂▂▂▂
Scale	2010	12.66	2010	2017	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▃▃▂▂▂▂▂▂
Unmanned aircraft	2010	4.54	2010	2017	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▃▃▂▂▂▂▂▂
Accuracy	2011	8.77	2011	2017	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▃▂▂▂▂▂▂
Camera	2011	8.73	2011	2018	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▃▃▂▂▂▂▂
Classification	2011	8.34	2011	2016	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▂▂▂▂▂▂▂
Chlorophyll fluorescence	2014	5.31	2014	2018	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▂▂▂▂▂
Radiometric calibration	2014	4.82	2014	2019	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▂▂▂▂
Hyperspectral imagery	2014	3.96	2014	2019	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▃▂▂▂▂
Structure from motion	2014	8.06	2015	2019	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▂▂▂▂
Fluorescence	2015	6.29	2015	2019	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▃▂▂▂▂
Photosynthesis	2015	5.86	2015	2018	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▂▂▂▂▂
Topography	2016	7.26	2016	2018	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▂▂▂▂▂
Calibration	2016	5.68	2016	2019	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▂▂▂▂
Point clouds	2016	5.47	2016	2017	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂▂▂▂▂
Digital surface model	2016	5.13	2016	2019	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▂▂▂▂
Digital elevation model	2016	4.07	2016	2019	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▂▂▂▂
High resolution	2016	3.77	2016	2019	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▂▂▂▂
Stomatal conductance	2017	4.76	2017	2020	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▂▂▂
Drought	2017	3.74	2017	2020	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▂▂▂
Coastal	2017	3.58	2017	2018	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂▂▂▂
Winter-wheat	2017	3.56	2017	2020	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃▃▂▂▂
Maize	2018	3.67	2018	2019	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂▂▂
Water quality	2020	4.25	2020	2021	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂
Infrared thermography	2020	3.82	2020	2021	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂
Individual tree detection	2020	3.82	2020	2021	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂
Land cover classification	2020	3.73	2020	2021	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▂▂
Vegetation mapping	2020	4.35	2021	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃
Precision	2021	4.51	2021	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃
Landsat 8	2021	3.91	2021	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃
Spatial variability	2021	3.91	2021	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃
Evaporation	2021	3.61	2021	2024	▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▂▃▃▃

). Since entering the scientific landscape in the early 21st century, UAV-RS technology has undergone four major phases: platform diversification, sensor integration, intelligent data processing, and real-time application. Initially, a variety of flight platforms—multi-rotor, fixed-wing, and unmanned helicopters—emerged, leveraging high-precision satellite navigation and inertial navigation systems to achieve breakthroughs in long-endurance flights, high-precision autonomous positioning, and vision-assisted navigation [134,135,136,137]. These advancements, coupled with integrated obstacle avoidance technologies combining LiDAR and visual sensors, significantly enhanced flight safety [138,139,140,141]. Subsequently, progress in miniaturization and diversification of sensors accelerated. Optical cameras evolved into multispectral, hyperspectral, and thermal infrared systems, while lightweight LiDAR systems enabled UAVs to capture both high-resolution 2D imagery and 3D point cloud data [142,143,144]. To facilitate the fusion of heterogeneous datasets, sensor calibration algorithms and coordinate transformation methods were further refined, effectively addressing issues arising from varying resolutions and coordinate systems [145,146].

Since 2015, as UAV technology matured, “unmanned aerial vehicle” itself became a burst keyword, highlighting a shift toward intelligent data processing. Emphasis on data quality was reflected in emerging keywords such as “accuracy” and “LiDAR.” Machine learning-based preprocessing, radiometric calibration, and noise suppression significantly improved data quality at the acquisition stage. Deep learning frameworks achieved major advances in object classification, semantic segmentation, and feature extraction, enabling precise recognition in complex environments [147,148,149]. The integration of SfM 3D reconstruction with change detection algorithms expanded the capacity for large-scale spatiotemporal analysis [150,151,152]. Keywords like “precision agriculture” and “classification” emerged as frontier topics, underscoring the explosion in application-driven research. The most recent burst term, “vegetation” (from 2020 to present), reaffirms the central role of ecological and agricultural monitoring as primary drivers in the field.

In recent years, real-time and edge processing capabilities have gained momentum. With the integration of edge computing and efficient wireless communication, UAVs can now perform preliminary image interpretation and data alerts on-site, supporting water quality monitoring [153], crop growth assessment [154], wetland ecosystem surveillance [14], and disaster response [155]. This evolution not only reduces data transmission latency significantly, but also lays the foundation for a collaborative “UAV + cloud + user-end” operational model.

4. Core Insights and Critical Reflections

4.1. Evolution and Breakthroughs of Key Technologies

The scientific evolution path of UAV-RS is profoundly reflected in the continuous evolution and breakthroughs of its key technologies, which mainly include the core aspects of flight platforms, sensor technology, flight control and navigation systems, as well as data processing and analysis methods (

Table 8. Evolution and breakthroughs of key technologies in UAV-RS.

Field of Technology	Evolution and Breakthrough	Representative Technologies/Trends
Flight Platform	From single-function to diversification and high-performance; progress in materials, aerodynamic configuration, and navigation and control technologies.	Mult-rotor, fixed-wing, and VTOL UAVs; lightweight and high-strength materials such as carbon fiber; long-endurance, high-payload, strong environmental adaptability, and high-degree autonomy.
Sensor Technology	From single-optics to multi-type integration; miniaturization, light-weight, high-precision, intelligentization, and integration.	RGB, multispectral, hyperspectral, thermal infrared, LiDAR, SAR, and gas sensors; solid-state LiDAR; polarimetric remote sensing; integration of sensors with AI.
Flight Control and Navigation Systems	From manual remote control to high-degree autonomy; significant improvement in positioning and attitude control accuracy; intelligent path planning and obstacle avoidance.	GNSS (GPS, Beidou), IMU, RTK/PPK, SLAM, VIO; AI-assisted navigation (TAN); UAV swarm; advanced guidance, navigation, and control (GNC).
Data Processing and Analysis Methods	From traditional photogrammetry to intelligent processing; widespread application of deep learning; driven by the demand for real-time and automated processing.	SfM-MVS, radiometric/geometric correction, image fusion/classification; deep learning (CNN, RNN, GAN) for classification, object detection, segmentation, change detection; edge computing; cloud computing; big data processing technologies.

). The progress of these technologies collectively drives the enhancement of UAV-RS capabilities and the expansion of its application scope.

4.1.1. Evolution of Flight Platforms

The UAV flight platform is the foundation for the development of UAV-RS. Early UAV platforms had relatively single functions, with limited stability, payload capacity, and endurance. However, with the progress in material science, aerodynamics, and navigation and control technology, UAV platforms have shown a development trend towards diversification and high performance. The UAV-RS hardware platform mainly consists of two parts: the flight platform and the sensors it carries [23]. Among them, Multi-rotor UAVs have become the most popular platform type for near-ground remote sensing applications due to their high safety, easy operation, vertical take-off and landing capability, and adaptability in complex environments. Common configurations such as quadrotor, hexarotor, and octarotor do not require large airports or runways, can flexibly adjust flight altitude and speed, and some high-end models even have obstacle detection and avoidance functions. However, multi-rotor UAVs still have certain limitations in terms of endurance and payload capacity. For example, after careful optimization, consumer-grade UAVs usually have an endurance of about 30 min. In contrast, Fixed-wing UAVs and Unmanned Helicopters as well as VTOL UAVs (Vertical Take-Off and Landing UAVs) have advantages in endurance, flight speed, and coverage, but their take-off and landing conditions and requirements for operational skills are relatively high.

With the continuous optimization of flight platforms, it has become possible to carry more advanced sensors and perform more complex remote sensing tasks. These new types of platforms have many unique advantages, such as a flight altitude lower than the traditional aviation navigation limit, thus obtaining higher spatial resolution; low flight speed or even zero-speed hovering capability (for multi-rotor); the ability to fly in complex environments, such as between buildings or indoor mapping; and the use of lightweight and high-strength materials like carbon fiber. These advances are thanks to the development of electronic technology, including the Inertial Measurement Unit (IMU) for navigation and stabilization, the Global Navigation Satellite System (GNSS) for positioning and navigation, and efficient ground stations and control systems. However, civilian small UAVs still have problems such as sensitivity to wind and limited endurance, which restrict their coverage, payload, and mission time. Therefore, the improvement of energy storage systems, such as the application of new materials like photovoltaic cells, fuel cells (hydrogen energy), and graphene, is a current research hotspot [156].

4.1.2. Sensor Technology Innovation

The innovation of sensor technology is the key to enhancing the capability of UAV-RS information acquisition. A significant advantage of UAV-RS over satellite remote sensing is its flexible sensor configuration ability. It can carry different types of sensors according to mission requirements, thereby obtaining diversified ground target information [24]. Early UAV-RS mainly relied on visible light cameras to acquire RGB images. With the development of technology, advanced sensors such as Multispectral, Hyperspectral, Thermal Infrared, LiDAR, and Synthetic Aperture Radar (SAR) have gradually been integrated into UAV platforms. Multispectral and hyperspectral sensors can acquire spectral information of ground objects in different bands, providing a data basis for refined ground object classification, vegetation health status assessment, mineral identification, etc. For example, in crop pest and disease monitoring, multispectral data are the most widely studied type of data. Thermal infrared sensors can detect the temperature distribution of ground objects and have unique advantages in fields such as forest fire monitoring, urban heat island effect research, and power line inspection. LiDAR measures distance accurately by emitting laser pulses and receiving echoes, and can quickly obtain high-precision 3D point cloud data for topographic mapping, forest structure parameter inversion, 3D building modeling, etc. The emergence of Solid-state LIDAR technology, with its more compact and lightweight characteristics, is particularly suitable for UAV platforms. The continuous miniaturization, light-weighting, high-precision, and low-cost of sensor technology have greatly expanded the depth and breadth of UAV-RS applications. Polarization remote sensing, as an emerging direction, acquires multi-dimensional information such as the degree of polarization and polarization azimuth angle of ground objects. It shows great potential in sky polarization navigation, image dehazing, rock density inversion, etc. The Hefei Institute of Optics and Fine Mechanics of the Chinese Academy of Sciences and Peking University have made important progress in the development of UAV-borne polarization CCD cameras [157].

4.1.3. Advancements in Flight Control and Navigation Systems

Early UAV flight control mainly relied on manual remote control or simple autopilots. With the development of Global Navigation Satellite System (GNSS, such as GPS, GLONASS, Beidou) and Inertial Measurement Unit (IM)U technology, the positioning and attitude control accuracy of UAVs has been greatly improved. RTK (Real-Time Kinematic) technology can provide centimeter-level and even millimeter-level real-time dynamic positioning, which significantly improves the geometric accuracy of UAV-RS data. Flight control systems are also becoming increasingly intelligent, with functions such as autonomous path planning, obstacle avoidance, and adaptive flight. For example, the construction of a general physical model for UAV-RS systems has transformed the multi-rigid-body stitching of imaging payloads into a single rigid-body imaging method. This has achieved simple automatic control of the payload and, based on this, an integrated UAV-RS system has been built to realize automated dynamic remote sensing control and observation [157]. The progress of path planning algorithms, especially path planning methods based on machine learning and deep learning, such as Deep Q-Network (DQN), Convolutional Neural Network (CNN) path planning, and Generative Adversarial Network (GAN) path planning, enables UAVs to better adapt to dynamic environments, optimize flight paths, and improve the efficiency and safety of mission execution.

In the future, flight control and navigation systems will place greater emphasis on autonomy, robustness, and collaboration, such as multi-UAV collaborative operations and autonomous obstacle avoidance path and planning in complex environments. For example, an AI-assisted Terrain Aided Navigation (TAN) system can combine high-performance terrain servers and digital elevation models to achieve low-altitude navigation in complex terrain [158]. In addition, UAV swarm technology has also made significant progress. Through advanced Guidance, Navigation, and Control (GNC) technology and swarm intelligence software, multiple UAVs can autonomously collaborate under manned aircraft command to perform complex tasks such as real-time surveillance and threat detection, demonstrating great potential in Manned–Unmanned Teaming (MUM-T) [159].

4.1.4. Intelligentization of Data Processing and Analysis Methods

The automation, intelligence, and real-time processing of data are important directions for future development, aiming to improve data processing efficiency and shorten the time from data acquisition to information extraction to meet the needs of applications with high timeliness requirements, such as emergency response. UAV-RS can obtain a large amount of high-resolution image data, which puts forward higher requirements for data processing and analysis methods. Traditional remote sensing data processing methods, such as radiometric correction, geometric correction, image fusion, and classification identification, have been widely applied and developed in the field of UAV-RS. In recent years, with the rapid development of artificial intelligence technology, especially the rise of deep learning methods, revolutionary breakthroughs have been brought to UAV-RS data processing. In the papers published in recent years, the processing of image data (RGB, multispectral, hyperspectral) mainly adopts neural network methods. For example, algorithms for object detection, image segmentation, and scene classification based on Convolutional Neural Networks (CNN) have shown superior performance in land use/land cover mapping, change detection, disaster assessment, and other aspects. However, for some special types of data, such as LiDAR data, there is still a lack of end-to-end neural network processing methods [23]. In addition, SfM and Multi-View Stereo (MVS) technologies, which reconstruct three-dimensional scenes from overlapping sequential images to quickly generate high-precision Digital Surface Models (DSM) and 3D point clouds, have become an important means for UAV-RS to obtain elevation information. Moreover, the development of UAV-RS big data cloud processing technology has provided an effective solution for the storage, management, and analysis of massive data, promoting the development of UAV-RS towards intelligence and real-time processing.

4.2. Expansion and Deepening of Application Fields

The application fields of UAV-RS technology are constantly expanding and deepening with the evolution and breakthroughs of its key technologies. Initially dominated by military reconnaissance, UAV-RS has gradually penetrated into all aspects of civilian fields and is playing an increasingly important role in many industries (

Table 9. Expansion and deepening of the main application fields of UAV-RS.

Application Fields	Main Application Contents	Key Technologies/Sensors
Military Reconnaissance and Security Surveillance	Intelligence collection, battlefield surveillance, target location, damage assessment, border patrol, counter-terrorism operations, inspection of critical infrastructure, and security for large-scale events.	High-definition cameras, thermal infrared sensors, Synthetic Aperture Radar (SAR), and signals intelligence equipment.
Precision Agriculture	Crop growth monitoring, pest and disease identification and control, yield estimation, variable fertilization and irrigation, farmland environment monitoring, and crop phenotyping analysis.	Multispectral/hyperspectral sensors, thermal infrared sensors, RGB cameras; deep learning.
Forestry and Pasture Management	Forest resource investigation, tree species identification, forest fire monitoring and assessment, pest and disease monitoring, wildlife habitat investigation, grassland productivity assessment, and biomass estimation.	LiDAR, multispectral/hyperspectral sensors, RGB cameras.
Environmental Monitoring and Assessment	Atmospheric pollution monitoring, water body eutrophication and black-odorous water body monitoring, soil erosion and desertification investigation, wetland ecosystem monitoring, biodiversity investigation, coastal zone erosion monitoring, and oil spill accident emergency monitoring.	Multispectral/hyperspectral sensors, thermal infrared sensors, gas sensors (DOAS), LiDAR.
Disaster Emergency Response and Management	Rapid disaster assessment, mapping of the affected area, personnel search and rescue, investigation of lifeline engineering damage, and secondary disaster monitoring and early warning.	High-definition cameras, thermal infrared sensors, LiDAR, SAR.
Surveying and Three-Dimensional Modeling	Topographic mapping, cadastral surveying, engineering reconnaissance, mine monitoring, archaeological excavation, urban 3D modeling, and DSM generation.	RGB cameras, LiDAR; SfM-MVS technology.
Power Line Inspection	Refined inspection of power transmission lines, towers, insulators, etc., to promptly identify potential hidden dangers.	Visible light cameras, infrared/ultraviolet sensors, LiDAR.
Urban Planning and Management	Land use change monitoring, illegal construction identification, traffic flow monitoring, urban greening assessment, smart city construction, and municipal facility inspection.	RGB cameras, LiDAR; Oblique photogrammetry.
Other Emerging Applications	Mining (reserve estimation, mining monitoring), construction (construction progress monitoring, quality inspection), archaeology (site discovery, surveying), logistics and transportation (cargo delivery), low-altitude tourism.	Diversified sensor combinations.

). Its high flexibility, high resolution, low cost, and rapid response capabilities make it a powerful complement to traditional aerial and space remote sensing, and even show irreplaceable advantages in some specific applications. The robust growth of the UAV-RS market also reflects the extensive expansion and deepening of its application fields.

(1) Military reconnaissance and security surveillance are one of the earliest and most important application fields of UAV-RS. UAVs are capable of penetrating into enemy territory or high-risk environments to conduct intelligence collection, battlefield surveillance, target location, and damage assessment, effectively reducing casualties. For example, the United States Coast Guard has routinely deployed UAVs equipped with ISR (Intelligence, Surveillance, and Reconnaissance) sensors on National Security Cutters to build maritime situational awareness and support law enforcement missions and search and rescue operations [160]. Small UAVs like the “Black Hornet” are also used for military reconnaissance, providing local situational awareness for ground troops [161].

(2) Precision agriculture is one of the most active and mature fields of application for UAV-RS technology. UAVs can be equipped with multispectral and hyperspectral sensors to quickly obtain key agronomic parameters such as crop growth conditions, Leaf Area Index (LAI), and Above-Ground Biomass (AGB), providing a scientific basis for precision fertilization, variable rate pesticide application, pest and disease monitoring, and yield estimation [162]. For example, using UAV-RS data for crop pest and disease monitoring can quickly detect and locate diseased areas, guide precision pesticide application, and reduce the use of pesticides and environmental pollution [23]. In precision agriculture scenarios, the demand for real-time object detection is also increasing [163].

(3) In terms of forestry and rangeland remote sensing, UAVs can be used for forest resource surveys (such as tree species identification, tree height, diameter at breast height, and volume estimation), forest fire monitoring and assessment, pest and disease monitoring, wildlife habitat surveys, and rangeland productivity evaluation [22]. LiDAR sensors can penetrate the forest canopy to accurately obtain the terrain beneath the forest and individual tree structure parameters. High-resolution optical imagery and multispectral data can be used to identify tree species and monitor forest health conditions.

(4) Environmental monitoring and assessment is another important direction for UAV-RS applications. UAVs can be used for atmospheric pollution monitoring, such as carrying Differential Optical Absorption Spectroscopy (DOAS) sensors to monitor the emission of pollutants like NO₂ in industrial areas [164]. In terms of water environment monitoring, UAVs can quickly obtain information on water color, transparency, chlorophyll concentration, etc., to assess water quality and monitor algal blooms such as red tides [8]. In ecological surveys, UAVs can be used for wildlife population surveys, habitat assessment, and monitoring of special ecosystems such as wetlands and peatlands [125].

(5) The demand for UAV-RS technology in disaster emergency response and management is becoming increasingly urgent. After natural disasters such as earthquakes, floods, landslides, and debris flows, UAVs can quickly enter the disaster areas to obtain high-resolution disaster images, assess the scope of the disaster and the degree of damage, and provide timely information for rescue decision-making and material allocation [125]. For example, after an earthquake, UAVs can quickly generate a 3D model of the disaster area to help rescuers understand the situation of building collapses and plan rescue routes. In forest fire monitoring, UAVs can be equipped with thermal infrared sensors to monitor the spread of the fire in real time and provide support for firefighting command [22].

(6) The fields of surveying and mapping, geographic information acquisition, power line inspection, oil and gas pipeline monitoring, archaeological survey, and traffic monitoring also show great application potential [134,162,165]. For example, in the surveying and mapping industry, UAVs can quickly obtain high-precision topographic maps and 3D models, which greatly improves the efficiency of surveying and mapping and reduces costs. In the power industry, UAVs can be equipped with visible light and thermal infrared sensors to inspect transmission lines and promptly identify potential safety hazards. With the continuous progress of technology and the deepening of application needs, the application fields of UAV-RS will further expand, providing more precise, efficient, and safe remote sensing information services for various industries.

4.3. Challenges Facing UAV-RS

In terms of technology, UAV-RS still faces a series of challenges, which limit its performance in some specific application scenarios and wider promotion. The main technical bottlenecks include the real-time and accurate processing of data, the balance between sensor performance and cost, the endurance and stability of the flight platform, and the ability to operate in environments complex (

Table 10. Core bottlenecks of remote UAV sensing technology.

Technological Bottlenecks	Specific Manifestations	Impact and Challenges
Real-time and accurate data processing	Massive data volumes lead to time-consuming processing; complex algorithms (e.g., SfM, deep learning) involve high computational costs; edge computing capabilities are limited; and factors such as sensor calibration, illumination conditions, and algorithm robustness affect accuracy.	Meeting the requirements of emergency response and real-time monitoring remains challenging; data quality significantly impacts decision-making reliability; and problems such as variations in image orientation, high overlap ratios, and variable scales further complicate processing.
Sensor performance and cost	High-performance sensors (hyperspectral, LiDAR) are expensive; miniaturization, lightweight, and low power consumption still have room for improvement; calibration and maintenance require high professionalism; consumer-grade sensors have limited performance.	Limits its popularity among low-cost platforms and small- and medium-sized users; affects data quality and application effects; some sensors are difficult to meet specific application requirements (such as detection inside the canopy).
Endurance and stability of flight platforms	Battery technology bottlenecks lead to short flight time; poor flight stability under complex weather conditions; limited payload capacity; and impacted by electromagnetic interference.	Limit the scope and efficiency of operations; affect data acquisition quality and flight safety; and make it difficult to undertake large-area, long-term monitoring tasks.
Ability to operate in complex environments	Navigation positioning accuracy decreases in urban canyons, dense forests, and other environments; communication links are easily interfered or blocked; special terrain or lighting conditions affect imaging quality.	Limits its application in complex environments; affects the reliability and integrity of data acquisition.
Data management, storage, and sharing	Massive data places high demands on storage and management; the lack of unified data standards and formats makes sharing and interoperability difficult.	Increase data management costs; limit the comprehensive utilization and value mining of data.

).

4.3.1. Platform Performance Limitations

Despite the notable advantages of UAV-RS in platform flexibility, data precision, and low-altitude observation capabilities, its practical deployment still encounters multiple technical and engineering challenges. This study further identifies four core bottlenecks that restrict the large-scale, high-frequency, and high-precision application of UAVs:

(1)

Limited Flight Time. For electric UAV platforms, most commercial and industrial-grade drones currently support flight durations of only 30–60 min, constrained by the energy density limits of existing battery technologies [166,167,168]. In large-scale or long-term monitoring tasks, the need for frequent returns to replace batteries not only slows mission progress, but also increases safety risks, especially in remote areas or under extreme weather conditions [169]. Insufficient flight time directly affects spatial coverage and mission depth [170]; for instance, when monitoring large-scale regions at high frequencies, limited endurance significantly hampers continuous observation efficiency. To address this, researchers are exploring advanced energy solutions such as lithium-sulfur batteries, hydrogen fuel cells, and hybrid power systems to improve energy density and cycle life [171,172]. Additionally, flight paths and mission planning are being optimized using genetic algorithms and reinforcement learning to maximize battery efficiency and reduce unnecessary energy consumption [173]. For example, in agricultural monitoring—such as pest infestation assessment—UAV flight routes must account for wind conditions and light availability. However, strong winds or rainy weather can prevent takeoff and landing, further exacerbating the risk of mission disruption due to limited endurance. Currently, research has evolved from single-pass, single-site monitoring to multi-pass, multi-site continuous monitoring. While this enhances monitoring frequency, it also significantly increases operational workload and charging demands [174], placing greater pressure on backend data processing workflows [175].

(2)

Limited Payload Capacity. High-resolution optical, multispectral, thermal infrared, and LiDAR sensors are essential for detailed surface monitoring. However, these sensors are often bulky and heavy, placing strict demands on the payload capacity of small UAVs [176]. Currently, lightweight UAVs can typically only carry basic optical cameras, making it difficult to simultaneously conduct multi-modal and multi-angle data acquisition tasks [177]. Although larger UAV platforms are capable of supporting combinations of advanced sensors (e.g., multispectral + LiDAR + thermal infrared), their high cost, operational complexity, and stricter requirements for takeoff and landing areas limit their application in urban or confined environments [176]. While modular designs have improved the flexibility of sensor replacement and integration [178], they have not fundamentally resolved the contradiction between payload capacity and energy constraints. Overall, limited payload capacity remains one of the key challenges for UAV-RS systems, especially in complex monitoring scenarios. Nonetheless, with ongoing technological advancements and improvements in UAV design, this limitation is expected to be progressively mitigated.

(3)

Sensor Performance Bottlenecks. The resolution, sensitivity, field of view, and spectral range of sensors directly determine the quality of remote sensing data. High-resolution sensors can capture individual plants or subtle thermal anomalies, but they often require higher sampling rates and larger data bandwidths, posing challenges for UAV endurance and onboard storage capacity [179,180,181,182]. Infrared and thermal imaging are indispensable for applications like wildfire monitoring, but typically have lower spatial resolution than visible-spectrum imaging and are more susceptible to interference from clouds and fog [183,184]. Moreover, the trade-off between field of view and resolution makes it difficult for operators to balance coverage area with detail acquisition, increasing the complexity of flight path design and operational costs. While current technologies face limitations in resolution, viewing angle, and sensitivity, ongoing advances in sensor development and UAV payload capacity are expected to enable future remote sensing missions to collect multi-dimensional environmental data with greater efficiency and precision.

(4)

Limited Data Transmission Rates. In applications such as real-time monitoring and emergency response, UAVs are required to transmit large volumes of high-definition imagery, multispectral data, and video streams back to ground stations or cloud platforms promptly. While current Wi-Fi and 4G networks can support basic data transfer in urban areas, they often fail to meet bandwidth requirements in remote or low-signal regions [185,186], forcing UAVs to store data onboard and upload it only after returning, resulting in significant latency. Although satellite communication and enhanced wireless transmission technologies can offer broader coverage in harsh environments [187], their high costs and limited accessibility hinder widespread adoption. However, the advancement of 5G and next-generation satellite networks, along with improvements in onboard AI edge computing, is expected to enable real-time preprocessing and compression on UAV platforms, thereby alleviating the burden of large-scale data transmission in the future.

4.3.2. Data Processing and Analysis Challenges

UAV-RS technology has made continuous breakthroughs in high-resolution imaging, low-altitude flexible observation, and intelligent data processing. However, its large-scale deployment still faces deep-rooted data-related challenges. With the continuous improvement of sensor resolution and acquisition frequency, the volume of UAV-RS generated data is growing exponentially, resulting in an increasingly prominent issue of “data overload.” From data acquisition to transmission, storage, processing, integration, and sharing, each stage imposes greater demands on computational resources, algorithm efficiency, system architecture, and regulatory oversight.

(1) Massive Data Storage and Management. High-resolution optical, multispectral, and hyperspectral sensors often generate data measured in gigabytes (GB) or even terabytes (TB). Traditional storage systems and network bandwidth struggle to support real-time data transmission and processing, especially in remote monitoring scenarios [188,189,190]. The costs and complexity of data storage have significantly increased, making efficient compression and rapid retrieval essential while ensuring data integrity.

(2) Data Quality and Noise Suppression. Flight attitude jitter, illumination variation, sensor calibration errors, and adverse weather conditions introduce noise and image distortion, which undermine down-stream analysis accuracy [44,46]. Sensor discrepancies can lead to systematic biases [191,192,193], and improper exposure or white balance settings may reduce image quality [194,195]. Therefore, the development of adaptive radiometric calibration, motion compensation, and multi-source noise suppression algorithms is crucial.

(3) Massive Data Processing Pressure and Algorithm Complexity. UAV-RS generates data with high temporal frequency and spatial resolution—hundreds of GBs from a single flight. Traditional image processing methods are inadequate for automated, rapid response and large-scale analysis [196,197,198,199]. Although deep learning has shown promise in remote sensing, it requires large annotated datasets and high computational power [200], making it resource-intensive and not yet fully suitable for real-time field operations. This calls for more efficient algorithms and advanced hardware to handle large-scale remote sensing data in real time.

(4) Multi-Source Data Fusion and Heterogeneous Processing. Discrepancies in spatial and temporal resolution, coordinate systems, and data formats among optical, infrared, LiDAR, and SAR data increase the difficulty of geometric correction and fusion [201,202,203,204]. Processing heterogeneous data demands seamless integration across platforms, necessitating improved data standardization and protocol unification. Real-time applications like disaster monitoring require efficient multi-source fusion frameworks and parallel computing strategies [31,205,206], as well as enhanced quality assessment and uncertainty analysis [207,208,209]. The development of standardized multi-source remote sensing frameworks is imperative [24,199,200,201]. Solving these challenges requires interdisciplinary collaboration, novel algorithms and techniques, and the promotion of interoperability standards.

Additionally, the diversity of UAV-RS application scenarios and sector-specific data standards have led to data silos and redundant efforts. Unifying data formats, metadata specifications, and quality assessment procedures is essential for enabling cross-platform and cross-domain data sharing and reuse. Moreover, UAVs can easily collect personal and sensitive geospatial information, raising legal and ethical concerns [23,210,211]. Most countries’ legal frameworks remain underdeveloped, and UAV-RS use in urban areas lacks clear regulation [212,213,214,215,216]. It is therefore urgent to establish technical standards and legal frameworks covering data encryption, access control, and accountability.

4.4. Constraints of Policies, Regulations, and Standardization Systems

The rapid development of UAV-RS has also introduced a range of challenges related to policies, regulations, and standardization systems (

Table 11. Key constraints of the policy, regulatory, and standards framework for UAV-RS.

Constraints	Specific Manifestations	Impact and Challenges
Airspace management and flight approval	UAV operations are constrained by restricted airspace, altitude, and coverage; the approval procedures are often complex and time-intensive; regulatory frameworks differ across regions; and a unified, efficient air traffic management system is still lacking.	Such constraints reduce the flexibility and efficiency of operations, elevate associated costs, impede the implementation of large-scale and high-frequency missions, and weaken the capacity for timely emergency response.
Data security and privacy protection	High-resolution data may involve sensitive content, posing risks of leakage and misuse; the legal and regulatory framework remains underdeveloped, with a lack of clear operational guidelines and well-defined responsibilities.	These issues trigger public concerns, pose risks to personal privacy and corporate confidentiality, elevate data compliance challenges, and obstruct effective data sharing and utilization.
Lack of industry standards and regulations	Unified standards are lacking across various aspects, including UAV platforms, sensors, data formats, processing procedures, and the quality of resulting products.	Inconsistencies in data quality and formats hinder comparability and interoperability, increase the challenges for users in selecting and utilizing data, and ultimately impede industry-wide collaboration and technological advancement.
Improvement and adaptability of the regulatory system	Regulatory frameworks have not kept pace with technological advancements; legal provisions remain ambiguous or overly complex; and coordination among oversight bodies is inadequate.	Such challenges lead to confusion and compliance uncertainties for operators, constrain technological innovation and market development, and impede effective responses to emerging application needs.
Responsibility identification and insurance mechanism	There is a lack of clarity in determining accident liability, and the associated legal frameworks and insurance mechanisms remain inadequate.	Accidents often lead to difficulties in liability attribution and compensation, which in turn elevates the overall operational risk.

). These constraints not only affect the lawful and compliant operation of UAV-RS, but also hinder its application expansion and industrialization in specific fields. Key issues include airspace management, data security and privacy protection, and the absence of unified industry standards.

(1) Airspace management and flight approval are among the primary bottlenecks constraining the development of the low-altitude economy. Airspace is a limited resource, and how to plan and manage low-altitude airspace effectively to ensure the safe operation of UAVs alongside other aircraft remains a major challenge for governments and aviation authorities worldwide. Currently, many countries and regions impose strict regulations on UAV flight activities, such as designating no-fly zones, restricting flight altitude and range, and requiring pre-flight plan approvals [47]. While these measures aim to ensure aviation safety and national air defense, they also pose significant obstacles to UAV-based remote sensing operations. For example, it is often difficult to obtain flight approvals for large-scale UAV-RS missions, limiting their practical application. In the United States, the Federal Aviation Administration (FAA) requires UAVs to operate within the visual line of sight (VLOS) of the operator, mandates registration for commercial drones, and enforces compliance with Part 107 regulations, which stipulate that operations must be conducted by certified remote pilots [217]. In China, the use of low-altitude airspace is strictly regulated, and the proportion of accessible airspace for light and small UAVs (typically below 120 m) that does not require approval is relatively low. The lack of a flexible airspace access mechanism makes it difficult for UAVs to operate in densely populated areas or near critical infrastructure, thus limiting their applications in key domains such as urban management and emergency response.

(2) Data security and privacy protection have become growing concerns for both the public and regulatory authorities. Sensors mounted on UAVs—particularly high-resolution cameras—are capable of capturing vast amounts of imagery that may contain personal information, corporate data, or sensitive facility details. If mishandled or misused, such data could pose serious risks of data breaches and privacy violations [218]. For example, UAVs operating over urban areas may inadvertently capture private activities of residents or confidential business operations. How to regulate the collection, storage, transmission, use, and disposal of UAV-acquired data—while safeguarding personal privacy and commercial confidentiality—has become an urgent legal and ethical challenge. At present, the legal and regulatory frameworks concerning UAV data security and privacy protection remain underdeveloped, with a lack of clear operational guidelines and defined responsibilities. There is an urgent need to formulate dedicated management regulations for UAV-RS data, clarifying data ownership, usage rights, and privacy protection requirements, while also strengthening oversight and enforcement mechanisms.

(3) The lack of industry standards and specifications is another major bottleneck hindering the standardized development and widespread application of UAV-RS. As an emerging interdisciplinary field, UAV-RS currently lacks unified industry standards and technical specifications across many critical aspects. These include performance standards for UAV platforms, sensor calibration and data quality benchmarks, standardized data processing workflows and product specifications, and application service protocols. The absence of such standards has resulted in inconsistent data quality across different manufacturers and research institutions, making it difficult to compare and interoperate results, thus undermining data sharing and practical applications [219]. For instance, in Structure from Motion Multi-View Stereo (SfM-MVS) data processing, the use of “black-box” software packages often hinders error isolation or correction, as they rarely provide clearly interpretable bundle adjustment reports [220]. The lack of standardized data processing procedures and accuracy assessment methods also compromises the credibility and comparability of UAV-RS outputs. Establishing a comprehensive standard system for UAV-RS is essential for regulating market practices, ensuring data quality, and promoting the healthy development of the industry. This requires joint efforts from governments, industry associations, research institutions, and enterprises. For example, in the field of engineering surveying and mapping, the unification of technical standards remains one of the most urgent issues to address.

(4) Improving and Adapting the Regulatory Framework. With the rapid development of UAV technology and the continuous expansion of its application domains, existing regulatory systems often struggle to fully adapt to new circumstances. Regulatory policies may suffer from being outdated, ambiguous, or overly complex, thereby creating confusion and compliance risks for UAV operators [218]. For example, some regulations lack clear or flexible provisions regarding UAV definitions, classifications, operator qualifications, and operational licensing. Additionally, challenges remain in terms of coordination between regulatory bodies and the harmonization of cross-regional oversight. A dynamic and adaptive regulatory framework is urgently needed—one that keeps pace with technological advancements, clearly delineates regulatory responsibilities, simplifies approval processes, and strengthens enforcement against violations.

(5) Liability Attribution and Insurance Mechanisms. During flight or operational activities, UAVs may encounter unexpected accidents that result in personal injury or property damage. Determining liability—whether it lies with the operator, manufacturer, or a third party—and establishing corresponding insurance mechanisms are critical for safeguarding stakeholders’ interests and ensuring the sustainable development of the industry. At present, relevant laws, regulations, and insurance products remain underdeveloped. In the event of an accident, the attribution of responsibility and compensation may prove challenging. It is essential to clearly define the legal responsibilities associated with UAV operations and to promote the development of diversified UAV insurance products to help mitigate and distribute operational risks.

4.5. Regional Development Imbalance and Barriers to Technology Transfer

The global research capacity in UAV-RS is primarily concentrated in China, the United States, and several EU countries or regions. Low- and middle-income countries face significant barriers in equipment acquisition, talent development, and technological application, leading to imbalanced global collaboration and limited capacity for technology transfer and knowledge sharing. Although there are examples of international partnerships—such as collaboration between Florida State University and the Institute of Geographic Sciences and Natural Resources Research at the Chinese Academy of Sciences—the overall frequency and depth of international cooperation remain insufficient, constraining the diffusion of technologies and the sharing of best practices.

Furthermore, different countries prioritize research directions based on their domestic needs, resulting in fragmented outcomes. For instance, China emphasizes precision agriculture, the United States focuses on urban monitoring, while Europe concentrates on environmental protection [221]. This lack of a unified, coordinated strategy to address global challenges hinders technology transfer and broader application. Many of the most advanced technologies remain concentrated in a few high-income countries and are difficult to scale or implement in low- and middle-income regions [222,223]. This fragmentation not only limits the overall enhancement of global UAV-RS capabilities, but also constrains the technology’s contribution to the global sustainable development agenda. Core bottlenecks in UAV-RS platforms and data processing, coupled with asymmetrical global collaboration mechanisms, continue to restrict the widespread application and long-term innovation of UAV-RS technologies. Looking ahead, coordinated efforts are needed across several fronts—including the development of energy-efficient platforms, standardization of multi-source data, advancement of intelligent analytical algorithms, and the establishment of inclusive international cooperation frameworks—to fully unlock the potential of UAV-RS in Earth observation and environmental management.

4.6. Critical Commentary

4.6.1. Double-Edged Sword of Application-Driven Development: The Risk of Technological Stagnation Behind the Boom

This study quantitatively reveals the overwhelming dominance of an “application-driven” paradigm in the development of UAV-based remote sensing (UAV-RS), where progress is primarily propelled by real-world applications rather than technological innovation. The prominence of application-centered clusters (e.g., #0 vegetation indices) and the leading roles of institutions such as the Ministry of Agriculture and Rural Affairs of China and the USDA underscore a research agenda shaped by the immediate, practical needs of sectors like agriculture and environmental management. This pattern presents a double-edged sword. On one hand, application-driven research ensures close alignment with real-world demands, generating substantial and visible economic and societal benefits. This alignment is a key reason the field continues to receive funding and policy support. On the other hand, the overwhelming focus on applied outcomes risks the strategic neglect of foundational research, which is manifested in two major ways:

(1)

Persistent platform bottlenecks: Issues such as “endurance” and “battery life”—core limitations affecting UAV performance—have never emerged as prominent in burst keyword analysis, indicating a disconnect between academic priorities and industrial pain points. Yet, numerous technical reports and review papers have repeatedly emphasized endurance as the key constraint on large-scale, long-term UAV-RS applications [133].

(2)

Surface-level AI adoption: Although deep learning remains a trending topic, most high-frequency terms refer to the direct adoption of established computer vision models like YOLO and U-Net. There is a notable absence of original AI architectures tailored to the unique characteristics of remote sensing data, such as spatial context or multimodal inputs. This signals a risk of “imported convenience” taking precedence over “remote sensing-native” innovation, which may lead to homogenized applications and diminished technological differentiation [224,225].

If uncorrected, this trend may trap the UAV-RS field in a state of “high-level repetition,” where technologies are widely adopted but the core capabilities remain stagnant—ultimately limiting the potential to tackle more complex scientific challenges.

4.6.2. Data Integration Potential Risks: The Overlooked Foundation of Interoperability

Currently, UAV-mounted sensors exhibit significant differences in imaging mechanisms, spatial resolution, spectral composition, and data formats, posing substantial challenges for the integrated processing and efficient utilization of multi-source data. The data integration crisis has emerged as one of the most critical bottlenecks in the field. While the academic community has shown strong enthusiasm for multi-source data fusion, there is a notable disconnect between data fusion and “data standardization”. Keywords such as “fusion”, “multi-sensor”, and “LiDAR” appear frequently, reflecting an urgent demand to integrate heterogeneous data sources. However, practical and foundational terms like data standards, interoperability, OGC standards, and metadata are conspicuously absent from the list of high-frequency or high-burst keywords.

This reveals a fundamental flaw: the research community is often focused on ad hoc integrations while neglecting the construction of scalable and reproducible infrastructure necessary for sustainable data fusion. As a result, many research teams are building isolated “data silos” with solutions that are difficult to replicate, compare, or extend for secondary development. Thus, the primary challenge is not merely “data overload”, but a deeper “data integration potential risks” rooted in the lack of shared standards. This severely hinders the capacity for large-scale, cross-institutional, and cross-regional collaborative data analysis. Addressing this crisis is essential to unlock the full potential of UAV-RS data and to enable its meaningful application across disciplines and regions.

4.6.3. The Imbalanced Cost–Benefit Equation: SfM’s Dominance over LiDAR

Keyword co-occurrence analysis reveals a highly representative phenomenon: in the field of UAV-RS 3D modeling, SfM exhibits significantly higher frequency and network centrality than LiDAR. Although both are core technologies for 3D reconstruction, the disparity in their usage popularity reflects an underlying cost–benefit trade-off in current research. SfM enables image acquisition and point cloud generation using standard RGB cameras, offering clear advantages such as low cost, ease of operation, and broad adaptability [226]. These qualities have made SfM a preferred 3D modeling approach for small- to medium-sized research teams and ecological applications [227]. In contrast, while LiDAR provides higher vertical accuracy and superior penetration in complex terrain [228], its high equipment costs, complex system integration, and stringent platform and power requirements restrict its use to high-budget or precision-critical applications [229]. From the perspective of data processing, SfM workflows are relatively mature, with widely adopted software such as Pix4D and Agisoft Metashape that are easy to use and accessible to non-experts. In contrast, LiDAR data preprocessing, point cloud registration, and filtering typically demand greater technical expertise and currently lack standardized, open-access processing frameworks [230]. As a result, despite the unmatched technical capabilities of LiDAR, SfM continues to dominate due to its superior cost-effectiveness, leading to an imbalance in the current research ecosystem. Moving forward, the development of lightweight LiDAR systems and integrated SfM–LiDAR hybrid modeling frameworks will be essential to simultaneously enhance accuracy and reduce cost-helping to overcome key technical bottlenecks in UAV-based 3D modeling.

4.6.4. Challenges in the Global Collaboration Landscape: From Technological Silos to Global Synergy

The dominance of countries and regions such as China, the United States, and the European Union reflects the strength of their regional research ecosystems. At present, global collaboration in the UAV-RS field exhibits a “strong hubs, weak bridges” pattern: while China and the U.S. have formed dense internal research networks, the bridging connections between them remain disproportionately weak relative to their immense academic output. Although China and the United States lead in UAV-RS research productivity, our analysis also reveals significant geographic imbalances. In particular, countries in Africa, Southeast Asia, Latin America, and parts of South Asia show notably low levels of academic participation in this field. We identify three key reasons and characteristics underlying this disparity:

(1) Structural and institutional barriers, including limited research infrastructure, restricted access to high-end UAV platforms and sensors, lack of technical training opportunities, and high financial barriers to publishing in high-impact international journals. (2) Many countries face import restrictions, high equipment costs, and complex regulations, which make it difficult to acquire and maintain a stable UAV-RS operational system. Moreover, the absence of localized open-access remote sensing datasets and standardized workflows in these regions hampers their integration into the global knowledge system. This leads to a skewed research agenda—despite facing urgent environmental, agricultural, or disaster-monitoring challenges, these regions are underrepresented in global academic discourse. (3) Language barriers, limited access to English-dominated academic platforms, and a lack of targeted support from international funding agencies further exacerbate the issue. As a result, “knowledge blind spots” have emerged in the development of UAV-RS technologies, marginalizing the practical needs of developing countries.

This phenomenon of “technological silos” presents a serious challenge to scientific progress. Global issues such as climate change, food security, and biodiversity loss require transnational, seamless data sharing and methodological consensus. The current fragmented model of cooperation not only hinders joint efforts to address these challenges, but also widens the digital divide between countries and regions, limiting the global accessibility and equitable distribution of advanced technologies.

5. Research Gaps and Prospects

5.1. Data Ecosystem Development: From Ad Hoc Integration to a Standardized Data Framework

Lack of unified data standards, metadata specifications, and reliable data-sharing mechanisms often leads to the proliferation of “data silos.” How can we establish a UAV-RS data-sharing mechanism and technical framework that balances data sovereignty, commercial confidentiality, and the principles of open science? How can we define a set of open standards encompassing data acquisition, processing, and metadata to enable true interoperability across global platforms? Future efforts must prioritize enhancing the interoperability and integrated analysis capabilities of UAV-RS systems—this is a critical step toward advancing UAV-RS applications. Research must overcome key obstacles in integrating multi-source and heterogeneous data across access, registration, fusion, and modeling stages. To this end, two main avenues are proposed: standardization of data frameworks and the development of fusion algorithms. This dual approach—merging standardization with algorithmic innovation—is key to unlocking the full value of UAV-RS data in global environmental monitoring and decision-making.

(1) Standardized Data Interfaces and Metadata Management. To address inconsistencies in resolution, spectral bands, and coordinate systems across various sensors (optical, multispectral, thermal infrared, LiDAR, SAR), it is essential to establish unified data description and metadata standards (e.g., OGC SensorML, ISO 19156). This will enable plug-and-play access to different sensors and facilitate automated calibration and registration during preprocessing of multi-source data [145,146,231].

(2) Multi-Source Data Fusion Algorithms and Spatiotemporal Interpolation. Research should combine traditional pixel-level fusion methods (e.g., weighted averaging, principal component analysis) with feature-level fusion based on deep learning. Developing models such as Cross-Modal Transformers and Graph Neural Networks (GNNs) can enable synergistic extraction and complementary enhancement of multi-source information [201,202,203]. In terms of spatiotemporal continuity, time-series models (e.g., temporal convolutional networks, autoregressive models) can be used for data interpolation and forecasting, improving the timeliness and accuracy of dynamic monitoring.

5.2. Breakthroughs in Core Technology Platforms: Integration of Intelligent Flight Platforms and Edge Computing

Platform endurance, payload capacity, and autonomous navigation are currently the primary bottlenecks in UAV-based remote sensing (UAV-RS) platforms. As UAV-RS application scenarios grow increasingly complex and task demands escalate, the technology is shifting from a “flight-centered observational tool” to a “cooperative system driven by intelligent sensing and decision-making.” This transition highlights the urgent need to address the lag in fundamental platform technologies exposed by the “application-driven” evolution model. Focused innovation is essential in three core areas: platform performance, intelligent computing, and system integration.

(1) What is the optimal technological pathway for next-generation UAV power systems (e.g., solid-state batteries, hydrogen fuel cells) across diverse application scenarios, balancing energy density, safety, and cost? Future UAV platforms must simultaneously enhance payload capacity and flight endurance while reducing airframe weight and energy consumption. High specific energy batteries—such as solid-state lithium, lithium-sulfur, and hydrogen fuel cells—offer promising potential, but require breakthroughs in cost-effectiveness, reliability, and safety [171,172]. Meanwhile, airframe designs should adopt carbon-fiber composites and topology optimization to minimize structural weight and enhance wind resistance. Flight control systems will increasingly incorporate autonomous functions, such as visual–inertial fusion for intelligent localization and obstacle avoidance, along with online path replanning capabilities in unknown environments, enabling UAVs to operate more stably in complex terrain and dynamic conditions [137,138,139,140,141].

(2) How can intelligent platforms be designed to deeply integrate edge computing, enabling real-time onboard analysis and decision-making to mitigate data overload from the source? Embedding high-performance AI inference chips (e.g., NVIDIA Jetson series, Google Edge TPU) within UAVs is expected to shift computation from the cloud to the onboard system, allowing for in situ image preprocessing, object detection, and change monitoring. This migration alleviates transmission bandwidth and latency issues, enabling near-real-time decision-making [187]. In privacy-sensitive or bandwidth-constrained environments, federated learning frameworks can distribute model training and updates across multiple UAVs and ground servers, preserving data isolation while optimizing global model performance, thus enhancing the overall level of UAV-RS intelligence.

5.3. High-Precision 3D Reconstruction and Dynamic Phenological Monitoring

To achieve high-precision and dynamic perception of ecological environments, future UAV-RS research will increasingly advance toward 3D modeling and ecological process simulation. While traditional 2D imagery can capture land cover and changes, it often fails to adequately represent terrain, structural complexity, and ecological processes in regions with significant elevation variation, dense vegetation, or urban infrastructure. Consequently, researchers are actively exploring the integration of 3D reconstruction techniques with phenological modeling, promoting a transition in remote sensing from “seeing” to “understanding deeply and comprehensively.”

(1) Hybrid Modeling by Integrating SfM and LiDAR.

SfM offers low-cost and easily implemented 3D reconstruction in multi-view image scenarios, but its performance is limited by poor texture and variable lighting. By integrating lightweight LiDAR with point cloud and mesh data from SfM, researchers can improve the accuracy and robustness of 3D modeling through multi-scale ICP (Iterative Closest Point) algorithms and deep learning-based point cloud registration networks [143,150]. This hybrid modeling approach is particularly effective for high-precision spatiotemporal monitoring of vegetation canopies, riverbank deformation, and urban structural changes.

(2) Phenological Modeling and Ecological Response.

Real-time and continuous phenological monitoring relies on high-frequency remote sensing data. UAV-RS enables the acquisition of high-resolution vegetation indices (e.g., NDVI, SAVI) and land surface temperature, which can be used to construct surface phenology time series. When combined with meteorological variables and soil moisture data, time-series clustering and change-point detection algorithms can be employed to identify plant growth phases, stress events, and ecosystem response mechanisms [232,233]. This analytical framework is essential for assessing the impact of climate change on agricultural productivity and ecosystem health.

5.4. Applied Ethics and Global Collaboration

In the context of global UAV-RS applications, there remains a lack of coordinated mechanisms to address global challenges and an absence of clear ethical guidelines and privacy regulations regarding data collection. Key questions remain unresolved: How can we build and operate a transnational research center to collaboratively address issues such as climate change and food security, while mandating the sharing of data and methodologies? In urban precision management, how should we technically and legally define the boundary between public interest and personal privacy in UAV-RS data acquisition, and formulate enforceable regulations? To promote the sustainable global development and collaborative innovation of UAV-RS technology, it is imperative to establish a comprehensive standardization system and a multi-tiered international cooperation network.

On standardization, efforts should leverage existing Earth observation and sensor standards established by international organizations such as the FAO, CEOS (Committee on Earth Observation Satellites), and OGC (Open Geospatial Consortium). A full life-cycle UAV-RS technical protocol—covering data acquisition, processing, and sharing—should be developed [23,210,217]. This framework should include flight operation parameters, sensor calibration procedures, data formats and metadata standards, quality assessment metrics, and security and privacy management guidelines. Based on this foundation, industry associations, research institutions, and enterprises should jointly compile UAV-RS best practice manuals. These should provide end-to-end demonstration cases across typical application scenarios (e.g., precision agriculture, forest monitoring, flood assessment), enabling the standards to be adopted in technical evaluations, pilot projects, and commercial deployments.

In addition, when developing global collaboration networks—such as the design of global benchmark datasets, field experiment programs, and multinational cooperation frameworks—priority should be given to the inclusion of geographically diverse regions. This approach not only enhances the inclusiveness of UAV-RS technologies, but also lays the foundation for building a truly globally applicable scientific system. For example, the “OpenUAV” model could serve as a useful reference by aggregating UAV-RS datasets from representative ecological zones, agricultural regions, and urban environments across different continents, along with corresponding interpretation algorithms and application case studies.

Through this dual strategy of standardization and global cooperation, UAV-RS interoperability and data quality can be significantly enhanced. At the same time, disparities between countries in hardware accessibility, software development capabilities, and talent cultivation can be narrowed. This will foster equitable sharing of scientific and technological achievements in support of sustainable development and environmental governance. Ultimately, it will help build an open, fair, and sustainable UAV-RS global ecosystem, providing critical support for tackling climate change, ensuring food security, and improving disaster early warning and mitigation capabilities.

6. Conclusions

With the continuous advancement of remote sensing technologies and the rapid development of UAV platforms, unmanned aerial vehicle remote sensing (UAV-RS) has become a vital component of the global Earth observation system. Based on a systematic scientometric analysis of relevant literature from the Web of Science database, this study presents a diagnostic review of the global research landscape, paradigm evolution, and underlying challenges in the UAV-RS field. It goes beyond traditional descriptive reviews by offering deeper analytical insights. The main findings are as follows:

(1)

Since 2015, UAV-RS research has entered a phase of rapid growth. Research priorities have expanded from early focuses such as image acquisition and platform performance to more advanced areas, including multi-source data fusion, deep learning–based interpretation, 3D modeling, and real-time monitoring—reflecting a paradigm shift from “image acquisition” to “intelligent sensing.” The field is highly interdisciplinary, encompassing remote sensing, environmental science, geographic information systems (GIS), engineering, and agricultural sciences. Keyword clustering analysis highlights prominent research hotspots such as vegetation indices, 3D modeling, image recognition, deep learning, and ecological stress, characterized by a high degree of technological integration and application-driven development.

(2)

A key finding of this study is that while the UAV-RS field is experiencing exponential growth, it is also facing three major structural imbalances:

①

A dual-core but weakly coupled global collaboration structure: China and the United States are the two major innovation hubs, yet direct cooperation between them remains limited, raising concerns about the formation of technological silos. We suggest fostering innovative collaboration models, such as: establishing multilateral alliances centered on global challenges (e.g., climate change, biodiversity, food security); creating unified data-sharing protocols and interoperability standards; initiating cross-border mobility programs for early-career researchers and joint field observation projects to promote knowledge exchange and methodological synergy.

②

An application-heavy, foundation-light research pattern: The academic agenda is heavily skewed toward immediate application needs—such as precision agriculture—while foundational technologies critical for long-term development, including endurance, sensor calibration, and data standardization, are comparatively neglected. We recommend that research funding agencies rebalance priorities and support key technological directions such as: high-efficiency power systems (e.g., fuel cells, hybrid UAVs); ultralight, high-resolution multi-sensor integration (e.g., LiDAR + hyperspectral); AI-enhanced autonomous navigation and flight planning; and lightweight real-time processing architectures for edge applications. Long-term investment is also needed in open-source toolchains, multi-environment test platforms, and interdisciplinary research infrastructures to avoid shallow innovation and redundant development.

③

A cognitive gap in the data ecosystem: Although researchers show strong interest in multi-source data fusion, there is a systemic neglect of the necessary data standards and interoperability frameworks. We recommend that international remote sensing standardization bodies (e.g., OGC, ISO/TC 211) take the lead in developing open standards covering sensor calibration, spatial accuracy, and data exchange formats.

(3)

The unique contribution of this study lies in advancing scientometric analysis from a descriptive inquiry of “what is” to a diagnostic critique of “why it matters.” Through quantitative evidence, this study reveals the overwhelming dominance of application-driven research, the hidden impact of cost-effectiveness on technology adoption (e.g., SfM vs. LiDAR), and the notable absence of critical discourse around data standards and interoperability in the literature. This diagnostic approach highlights both the prosperity and the potential risks embedded within the current research paradigm, offering a critical perspective for understanding the future trajectory of UAV and remote sensing technologies.

This study used the Web of Science Core Collection as the sole data source, which is heavily dominated by English-language journals. As a result, gray literature, local reports, and conference papers from non-English-speaking countries or regions (e.g., Latin America and parts of Asia) may be underrepresented. Future research could incorporate multi-source data integration from Scopus, Google Scholar, CNKI, and preprint servers (e.g., arXiv). Additionally, book chapters, technical reports, patents, and preprints were excluded from this study, which may have led to the omission of early methodological prototypes or the latest technical developments. Future studies are encouraged to include a broader range of literature types to enhance coverage and completeness. Lastly, the strict use of search terms such as TS = “remote sensing” may have resulted in the exclusion of UAV-related studies that do not explicitly mention “remote sensing” in their titles, abstracts, or keywords. Future research may consider using semantic search, multilingual indexing, and topic modeling techniques to overcome the limitations of keyword-based retrieval.