Search Results (46)

Geographic object-based image analysis (GEOBIA) has demonstrated substantial utility in remote sensing tasks. However, its integration with deep learning remains largely confined to image-level classification. This is primarily due to the irregular shapes and fragmented boundaries of segmented objects, which limit its applicability in semantic segmentation. While convolutional neural networks (CNNs) excel at local feature extraction, they inherently struggle to capture long-range dependencies. In contrast, Transformer-based models are well suited for global context modeling but often lack fine-grained local detail. To overcome these limitations, we propose GOFENet (Geo-Object Feature Enhanced Network)—a hybrid semantic segmentation architecture that effectively fuses object-level priors into deep feature representations. GOFENet employs a dual-encoder design combining CNN and Swin Transformer architectures, enabling multi-scale feature fusion through skip connections to preserve both local and global semantics. An auxiliary branch incorporating cascaded atrous convolutions is introduced to inject information of segmented objects into the learning process. Furthermore, we develop a cross-channel selection module (CSM) for refined channel-wise attention, a feature enhancement module (FEM) to merge global and local representations, and a shallow–deep feature fusion module (SDFM) to integrate pixel- and object-level cues across scales. Experimental results on the GID and LoveDA datasets demonstrate that GOFENet achieves superior segmentation performance, with 66.02% mIoU and 51.92% mIoU, respectively. The model exhibits strong capability in delineating large-scale land cover features, producing sharper object boundaries and reducing classification noise, while preserving the integrity and discriminability of land cover categories. Full article

Image analysis is a valuable approach in a wide array of environmental applications. Mapping land cover categories depicted from satellite images enables the monitoring of landscape dynamics. Such a technique plays a key role for land management and predictive ecosystem modelling. Satellite-based mapping of environmental dynamics enables us to define factors that trigger these processes and are crucial for our understanding of Earth system processes. In this study, a reclassification scheme of image analysis was developed for mapping the adjusted categorisation of land cover types using multispectral remote sensing datasets and Geographic Resources Analysis Support System (GRASS) Geographic Information System (GIS) software. The data included four Landsat 8–9 satellite images on 2015, 2019, 2021 and 2023. The sequence of time series was used to determine land cover dynamics. The classification scheme consisting of 17 initial land cover classes was employed by logical workflow to extract 10 key land cover types of the coastal areas of Bab-el-Mandeb Strait, southern Red Sea. Special attention is placed to identify changes in the land categories regarding the thermal saline lake, Lake Assal, with fluctuating salinity and water levels. The methodology included the use of machine learning (ML) image analysis GRASS GIS modules ‘r.reclass’ for the reclassification of a raster map based on category values. Other modules included ‘r.random’, ‘r.learn.train’ and ‘r.learn.predict’ for gradient boosting ML classifier and ‘i.cluster’ and ‘i.maxlik’ for clustering and maximum-likelihood discriminant analysis. To reveal changes in the land cover categories around the Lake of Assal, this study uses ML and reclassification methods for image analysis. Auxiliary modules included ‘i.group’, ‘r.import’ and other GRASS GIS scripting techniques applied to Landsat image processing and for the identification of land cover variables. The results of image processing demonstrated annual fluctuations in the landscapes around the saline lake and changes in semi-arid and desert land cover types over Djibouti. The increase in the extent of semi-desert areas and the decrease in natural vegetation proved the processes of desertification of the arid environment in Djibouti caused by climate effects. The developed land cover maps provided information for assessing spatial–temporal changes in Djibouti. The proposed ML-based methodology using GRASS GIS can be employed for integrating techniques of image analysis for land management in other arid regions of Africa. Full article

Collaboration between autonomous vehicles and drones can enhance the efficiency and connectivity of three-dimensional transportation systems. When satellite signals are unavailable, vehicles can achieve accurate localization by matching rich ground environmental data to digital maps, simultaneously providing the auxiliary localization information for drones. However, conventional digital maps suffer from high construction costs, easy misalignment, and low localization accuracy. Thus, this paper proposes a visual point cloud map (VPCM) construction and matching localization for autonomous vehicles. We fuse multi-source information from vehicle-mounted sensors and the regional road network to establish the geographically high-precision VPCM. In the absence of satellite signals, we segment the prior VPCM on the road network based on real-time localization results, which accelerates matching speed and reduces mismatch probability. Simultaneously, by continuously introducing matching constraints of real-time point cloud and prior VPCM through improved iterative closest point matching method, the proposed solution can effectively suppress the drift error of the odometry and output accurate fusion localization results based on pose graph optimization theory. The experiments carried out on the KITTI datasets demonstrate the effectiveness of the proposed method, which can autonomously construct the high-precision prior VPCM. The localization strategy achieves sub-meter accuracy and reduces the average error per frame by 25.84% compared to similar methods. Subsequently, this method’s reusability and localization robustness under light condition changes and environment changes are verified using the campus dataset. Compared to the similar camera-based method, the matching success rate increased by 21.15%, and the average localization error decreased by 62.39%. Full article

National-scale forest inventories such as the Forest Inventory and Analysis (FIA) program in the United States are designed to provide data and estimates that meet target precision at the national and state levels. However, such design-based direct estimates are often not accurate at a smaller geographic scale due to the small sample size. Small area estimation (SAE) techniques provide precise estimates at small domains by borrowing strength from remotely sensed auxiliary information. This study combined the FIA direct estimates with gridded mean canopy heights derived from recently published Global Ecosystem Dynamics Investigation (GEDI) Level 3 data and Landsat data to improve county-level estimates of total and merchantable volume, aboveground biomass, and basal area in the states of Alabama and Mississippi, USA. Compared with the FIA direct estimates, the area-level SAE models reduced root mean square error for all variables of interest. The multi-state SAE models had a mean relative standard error of 0.67. In contrast, single-state models had relative standard errors of 0.54 and 0.59 for Alabama and Mississippi, respectively. Despite GEDI’s limited footprints, this study reveals its potential to reduce direct estimate errors at the sub-state level when combined with Landsat bands through the small area estimation technique. Full article

Recently, Time Series Remote Sensing Images (TSRSIs) have been proven to be a significant resource for land use/land cover (LULC) mapping. Deep learning methods perform well in managing and processing temporal dependencies and have shown remarkable advancements within this domain. Although deep learning methods have exhibited outstanding performance in classifying TSRSIs, they rely on enough labeled time series samples for effective training. Labeling data with a wide geographical range and a long time span is highly time-consuming and labor-intensive. Active learning (AL) is a promising method of selecting the most informative data for labeling to save human labeling efforts. It has been widely applied in the remote sensing community, except for the classification of TSRSIs. The main challenge of AL in TSRSI classification is dealing with the internal temporal dependencies within TSRSIs and evaluating the informativeness of unlabeled time series data. In this paper, we propose a data-driven active deep learning framework for TSRSI classification to address the problem of limited labeled time series samples. First, a temporal classifier for TSRSI classification tasks is designed. Next, we propose an effective active learning method to select informative time series samples for labeling, which considers representativeness and uncertainty. For representativeness, we use the K-shape method to cluster time series data. For uncertainty, we construct an auxiliary deep network to evaluate the uncertainty of unlabeled data. The features with rich temporal information in the classifier’s middle-hidden layers will be fed into the auxiliary deep network. Then, we define a new loss function with the aim of improving the deep model’s performance. Finally, the proposed method in this paper was verified on two TSRSI datasets. The results demonstrate a significant advantage of our method over other approaches to TSRSI. On the MUDS dataset, when the initial number of samples was 100 after our method selected and labeled 2000 samples, an accuracy improvement of 4.92% was achieved. On the DynamicEarthNet dataset, when the initial number of samples was 1000 after our method selected and labeled 2000 samples, an accuracy improvement of 7.81% was attained. On the PASTIS dataset, when the initial number of samples was 1000 after our method selected and labeled 2000 samples, an accuracy improvement of 4.89% was achieved. Our code is available in Data Availability Statement. Full article

Multi-scale electronic map tiles are important basic geographic information data, and an approach based on deep learning is being used to generate multi-scale map tiles. Although generative adversarial networks (GANs) have demonstrated great potential in single-scale electronic map tile generation, further research concerning multi-scale electronic map tile generation is needed to meet cartographic requirements. We designed a multi-scale electronic map tile generative adversarial network (MsM-GAN), which consisted of several GANs and could generate map tiles at different map scales sequentially. Road network data and building footprint data from OSM (Open Street Map) were used as auxiliary information to provide the MsM-GAN with cartographic knowledge about spatial shapes and spatial relationships when generating electronic map tiles from remote sensing images. The map objects which should be deleted or retained at the next map scale according to cartographic standards are encoded as auxiliary information in the MsM-GAN when generating electronic map tiles at smaller map scales. In addition, in order to ensure the consistency of the features learned by several GANs, the density maps constructed from specific map objects are used as global conditions in the MsM-GAN. A multi-scale map tile dataset was collected from MapWorld, and experiments on this dataset were conducted using the MsM-GAN. The results showed that compared to other image-to-image translation models (Pix2Pix and CycleGAN), the MsM-GAN shows average increases of 10.47% in PSNR and 9.92% in SSIM and has the minimum MSE values at all four map scales. The MsM-GAN also performs better in visual evaluation. In addition, several comparative experiments were completed to verify the effect of the proposed improvements. Full article

It is crucial for both traffic management organisations and individual commuters to be able to forecast traffic flows accurately. Graph neural networks made great strides in this field owing to their exceptional capacity to capture spatial correlations. However, existing approaches predominantly focus on local geographic correlations, ignoring cross-region interdependencies in a global context, which is insufficient to extract comprehensive semantic relationships, thereby limiting prediction accuracy. Additionally, most GCN-based models rely on pre-defined graphs and unchanging adjacency matrices to reflect the spatial relationships among node features, neglecting the dynamics of spatio-temporal features and leading to challenges in capturing the complexity and dynamic spatial dependencies in traffic data. To tackle these issues, this paper puts forward a fresh approach: a new self-supervised dynamic spatio-temporal graph convolutional network (SDSC) for traffic flow forecasting. The proposed SDSC model is a hierarchically structured graph–neural architecture that is intended to augment the representation of dynamic traffic patterns through a self-supervised learning paradigm. Specifically, a dynamic graph is created using a combination of temporal, spatial, and traffic data; then, a regional graph is constructed based on geographic correlation using clustering to capture cross-regional interdependencies. In the feature learning module, spatio-temporal correlations in traffic data are subjected to recursive extraction using dynamic graph convolution facilitated by Recurrent Neural Networks (RNNs). Furthermore, self-supervised learning is embedded within the network training process as an auxiliary task, with the objective of enhancing the prediction task by optimising the mutual information of the learned features across the two graph networks. The superior performance of the proposed SDSC model in comparison with SOTA approaches was confirmed by comprehensive experiments conducted on real road datasets, PeMSD4 and PeMSD8. These findings validate the efficacy of dynamic graph modelling and self-supervision tasks in improving the precision of traffic flow prediction. Full article

Snow detection is imperative in remote sensing for various applications, including climate change monitoring, water resources management, and disaster warning. Recognizing the limitations of current deep learning algorithms in cloud and snow boundary segmentation, as well as issues like detail snow information loss and mountainous snow omission, this paper presents a novel snow detection network based on Swin-Transformer and U-shaped dual-branch encoder structure with geographic information (SD-GeoSTUNet), aiming to address the above issues. Initially, the SD-GeoSTUNet incorporates the CNN branch and Swin-Transformer branch to extract features in parallel and the Feature Aggregation Module (FAM) is designed to facilitate the detail feature aggregation via two branches. Simultaneously, an Edge-enhanced Convolution (EeConv) is introduced to promote snow boundary contour extraction in the CNN branch. In particular, auxiliary geographic information, including altitude, longitude, latitude, slope, and aspect, is encoded in the Swin-Transformer branch to enhance snow detection in mountainous regions. Experiments conducted on Levir_CS, a large-scale cloud and snow dataset originating from Gaofen-1, demonstrate that SD-GeoSTUNet achieves optimal performance with the values of 78.08%, 85.07%, and 92.89% for $IoU\_s$, $F1\_s$, and $MPA$, respectively, leading to superior cloud and snow boundary segmentation and thin cloud and snow detection. Further, ablation experiments reveal that integrating slope and aspect information effectively alleviates the omission of snow detection in mountainous areas and significantly exhibits the best vision under complex terrain. The proposed model can be used for remote sensing data with geographic information to achieve more accurate snow extraction, which is conducive to promoting the research of hydrology and agriculture with different geospatial characteristics. Full article

Long-term exposure to high concentrations of fine particles can cause irreversible damage to people’s health. Therefore, it is of extreme significance to conduct large-scale continuous spatial fine particulate matter (PM_2.5) concentration prediction for air pollution prevention and control in China. The distribution of PM_2.5 ground monitoring stations in China is uneven with a larger number of stations in southeastern China, while the number of ground monitoring sites is also insufficient for air quality control. Remote sensing technology can obtain information quickly and macroscopically. Therefore, it is possible to predict PM_2.5 concentration based on multi-source remote sensing data. Our study took China as the research area, using the Pearson correlation coefficient and GeoDetector to select auxiliary variables. In addition, a long short-term memory neural network and random forest regression model were established for PM_2.5 concentration estimation. We finally selected the random forest regression model (R² = 0.93, RMSE = 4.59 μg m⁻³) as our prediction model by the model evaluation index. The PM_2.5 concentration distribution across China in 2021 was estimated, and then the influence factors of high-value regions were explored. It is clear that PM_2.5 concentration is not only related to the local geographical and meteorological conditions, but also closely related to economic and social development. Full article

Soil organic carbon (SOC) is widely recognized as an essential indicator of the quality of arable soils and the health of ecosystems. In addition, an accurate understanding of the spatial distribution of soil organic carbon content for precision digital agriculture is important. In this study, the spatial distribution of organic carbon in topsoil was determined using four common machine learning methods, namely the back-propagation neural network model (BPNN), random forest algorithm model (RF), geographically weighted regression model (GWR), and ordinary Kriging interpolation method (OK), with Helan County as the study area. The prediction accuracies of the four different models were compared in conjunction with multiple sources of auxiliary variables. The prediction accuracies for the four models were BPNN (MRE = 0.066, RMSE = 0.257) > RF (MRE = 0.186, RMSE = 3.320) > GWR (MRE = 0.193, RMSE = 3.595) > OK (MRE = 0.198, RMSE = 4.248). Moreover, the spatial distribution trends for the SOC content predicted with the four different models were similar: high in the western area and low in the eastern area of the study region. The BPNN model better handled the nonlinear relationship between the SOC content and multisource auxiliary variables and presented finer information for spatial differentiation. These results provide an important theoretical basis and data support to explore the spatial distribution trend for SOC content. Full article

The ability to obtain accurate precipitation data from various geographic locations is crucial for many applications. Various global products have been released in recent decades for estimating precipitation spatially and temporally. Nevertheless, it is extremely important to provide reliable and accurate products for estimating precipitation in a variety of environments. This is due to the complexity of topographic, climatic, and other factors. This study proposes a multi-product information combination for improving precipitation data accuracy based on a generalized regression neural network model using global and local optimization strategies. Firstly, the accuracy of ten global precipitation products from four different categories (satellite-based, gauge-corrected satellites, gauge-based, and reanalysis) was assessed using monthly precipitation data collected from 1896 gauge stations in Iran during 2003–2021. Secondly, to enhance the accuracy of the modeled precipitation products, the importance score of effective and auxiliary variables—such as elevation, the Enhanced Vegetation Index (EVI), the Land Surface Temperature (LST), the Soil Water Index (SWI), and interpolated precipitation maps—was assessed. Finally, a generalized regression neural network (GRNN) model with global and local optimization strategies was used to combine precipitation information from several products and auxiliary characteristics to produce precipitation data with high accuracy. Global precipitation products scored higher than interpolated precipitation products and surface characteristics. Furthermore, the importance score of the interpolated precipitation products was considerably higher than that of the surface characteristics. SWI, elevation, EVI, and LST scored 53%, 20%, 15%, and 12%, respectively, in terms of importance. The lowest RMSE values were associated with IMERGFinal, TRMM3B43, PERSIANN-CDR, ERA5, and GSMaP-Gauge. For precipitation estimation, these products had Kling–Gupta efficiency (KGE) values of 0.89, 0.86, 0.77, 0.78, and 0.60, respectively. The proposed GRNN-based precipitation product with a global (local) strategy showed RMSE and KGE values of 9.6 (8.5 mm/mo) and 0.92 (0.94), respectively, indicating higher accuracy. Generally, the accuracy of global precipitation products varies depending on climatic conditions. It was found that the proposed GRNN-derived precipitation product is more efficient under different climatic conditions than global precipitation products. Moreover, the local optimization strategy based on climatic classes outperformed the global optimization strategy. Full article

Effective utilization of airborne synthetic-aperture (Airborne SAR) imagery often requires precise location of each image pixel. Historically, the positioning of airborne SAR imagery either relies on the use of reliable reference points to determine the relative position of the image, or requires the precise motion information of the aircraft and the characteristics of the SAR data collection system as input to determine the absolute position of the image. However, for many applications, the accuracy of traditional positioning methods is not high due to the challenge in obtaining the accurate geographic positions of reliable reference points and the inaccuracy of the recorded aircraft motion information. This study introduces an airborne SAR image planar positioning approach based on the premise that the systematic positioning error of the dual-view airborne SAR images are relatively consistent. The suggested planar positioning method applies the positioning auxiliary parameters of the initial ground-range airborne SAR image to ascertain the transformation relationship between the target’s initial geographic position and pixel position, and it then uses the equivalent equation for the position of the homologue point to assess the systematic positioning error of the SAR image and determine the geographic position of a pixel in a digital SAR image. This approach has advantages over previous techniques in that it requires no precise geographic position information of the ground reference points, and on the basis of using the RD model to accomplish coarse positioning of four corners of SAR image, it no longer needs aircraft trajectory data. Tests were conducted using two airborne SAR images actually captured, and the experimental results indicate that the proposed method can achieve high precision planar positioning of dual-view airborne SAR images. Error sources are analyzed and recommendations are given to improve image positioning accuracy in future airborne SARs. Full article

(1) Background: Along with the maturity of smart cities, digital villages and smart villages are receiving more attention than ever before as the key to promote sustainable rural development. The Chinese government has made great efforts in promoting the digital development of villages in recent years, as evidenced by policies intensively introduced by the central and local governments, making China a typical representative country in the world. (2) Methods: This paper evaluates the performance and geographic pattern of rural digital development by the Geographic Information System (GIS) in Gansu, a less developed province in western China, and analyzes the driving mechanism of rural digital development using GeoDetector, providing a basis for spatial zoning and differentiated policy design for the construction, planning and management of digital villages based on the GE matrix. (3) Results: First, the development of digital villages shows a prominent geographical imbalance, with 79 counties divided into leader, follower and straggler levels. Second, digital villages show unsynchronized development in different dimensions, with the village facilities digitalization index in the lead and the village economy digitalization index lagging behind. Thirdly, the development of digital villages is characterized by significant spatial correlation and spillover effects, with cold and hot counties distributed in clusters, forming a “center-periphery” structure. Fourth, the factors show significant influence differentiation. They are classified into all-purpose, multifunctional and single-functional factors by their scope of action, and into key, important and auxiliary factors by their intensity of action. Fifth, the interaction and driving mechanism between different factors is quite complex, dominated by nonlinear enhancement and bifactor enhancement, and the synergistic effect of factor pairs helps increase the influence by 1–4 times. (4) Conclusions: It is suggested that the government develop differentiated policies for zoning planning and management based on the level of digital development of villages in combination with the factor influence and its driving mechanism and promote regional linkage and common development and governance through top-level design. Full article

Quantitative assessment of the effectiveness of ecological restoration provides timely feedback on restoration efforts, and helps to accurately understand the extent of restoration, while providing scientific support for optimizing restoration programs. In recent decades, the Wuliangsuhai watershed in China’s Inner Mongolia Autonomous Region has been affected by anthropogenic activities, resulting in an increasingly unbalanced ecological environment. In order to curb environmental degradation, the local government implemented the “mountain, water, forest, field, lake and grass ecological protection and restoration project of the Wuliangsuhai watershed” from 2018 to 2020. The project has been completed and there is an urgent need for remote sensing monitoring to aid in performance evaluation. We took the ecological protection and restoration area of the Wuliangsuhai watershed in China as the research object, applied multi-source remote sensing imagery and auxiliary data such as meteorology and geographic basic data, extracted information of each evaluation index before and after the implementation of this project, and used the entropy value method to determine the index weights to comprehensively evaluate the ecological restoration effect. The results showed that after the implementation of the ecological restoration project, the vegetation coverage was further improved, the effectiveness of desert management was obvious, soil and water conservation capacity was strengthened, the ecosystem became more stable, and the areas with good environment were mostly located in the central and eastern parts. A total of 37.86% of the areas had obvious ecological restoration effects, and all indicators were further improved. Among the main treatment areas, the restoration effect of the Wuliangsuhai water ecological restoration and biodiversity conservation area was the best. The restoration effect will be further accentuated over time. This study provides a scientific reference for the further management of the ecological environment in the watershed and can provide a reference for the evaluation of the ecological restoration effect in similar areas in the future. Full article

In this paper, an image analysis framework is formulated for Landsat-8 Operational Land Imager and Thermal Infrared Sensor (OLI/TIRS) scenes using the R programming language. The libraries of R are shown to be effective in remote sensing data processing tasks, such as classification using k-means clustering and computing the Normalized Difference Vegetation Index (NDVI). The data are processed using an integration of the RStoolbox, terra, raster, rgdal and auxiliary packages of R. The proposed approach to image processing using R is designed to exploit the parameters of image bands as cues to detect land cover types and vegetation parameters corresponding to the spectral reflectance of the objects represented on the Earth’s surface. Our method is effective at processing the time series of the images taken at various periods to monitor the landscape dynamics in the middle part of the Congo River basin, Democratic Republic of the Congo (DRC). Whereas previous approaches primarily used Geographic Information System (GIS) software, we proposed to explicitly use the scripting methods for satellite image analysis by applying the extended functionality of R. The application of scripts for geospatial data is an effective and robust method compared with the traditional approaches due to its high automation and machine-based graphical processing. The algorithms of the R libraries are adjusted to spatial operations, such as projections and transformations, object topology, classification and map algebra. The data include Landsat-8 OLI-TIRS covering the three regions along the Congo river, Bumba, Basoko and Kisangani, for the years 2013, 2015 and 2022. We also validate the performance of graphical data handling for cartographic visualization using R libraries for visualising changes in land cover types by k-means clustering and calculation of the NDVI for vegetation analysis. Full article