With the development of deep learning, convolutional neural networks (CNNs) have been successfully applied in the field of change detection in heterogeneous remote sensing (RS) images and achieved remarkable results. However, most of the existing methods of heterogeneous RS image change detection only extract deep features to realize the whole image transformation and ignore the description of the topological structure composed of the image texture, edge, and direction information. The occurrence of change often means that the topological structure of the ground object has changed. As a result, these algorithms severely limit the performance of change detection. To solve these problems, this paper proposes a new topology-coupling-based heterogeneous RS image change detection network (TSCNet). TSCNet transforms the feature space of heterogeneous images using an encoder–decoder structure and introduces wavelet transform, channel, and spatial attention mechanisms. The wavelet transform can obtain the details of each direction of the image and effectively capture the image’s texture features. Unnecessary features are suppressed by allocating more weight to areas of interest via channels and spatial attention mechanisms. As a result of the organic combination of a wavelet, channel attention mechanism, and spatial attention mechanism, the network can focus on the texture information of interest while suppressing the difference of images from different domains. On this basis, a bitemporal heterogeneous RS image change detection method based on the TSCNet framework is proposed. The experimental results on three public heterogeneous RS image change detection datasets demonstrate that the proposed change detection framework achieves significant improvements over the state-of-the-art methods. Full article

Clarifying the evolution pattern of groundwater storage (GWS) is crucial for exploring the amount of available water resources at a regional or basin scale. Currently, the groundwater resources of Northeast China have been extensively exploited, but only limited studies have assessed the extent of GWS depletion and its driving mechanisms. In this study, the groundwater storage anomaly (GWSA) in the black soil region of Northeast China was explored based on the Gravity Recovery and Climate Experiment (GRACE) satellite combined with the Global Land Data Assimilation System (GLDAS) hydrological model. The results show that from 2002 to 2021, the overall GWSA decreased (−0.4204 cm/a), and specifically, the average rates of decrease in Heilongjiang, Jilin, and Liaoning Provinces were −0.2786, −0.5923, and −0.6694 cm/a, respectively, with the eastern, southern, and central parts of Heilongjiang, Jilin, and Liaoning Provinces losing seriously. Especially the GWSA deficit trend can reach −0.7471 cm/a in southern Jilin Province. The GWSA deficits in the three provinces from April to September were greater than 0.40 cm/a, while the deficit values from January to March and from October to December were less than 0.40 cm/a. This study is the first to quantitatively analyze the GWSA and its influencing factors in Northeast China for 2002–2021. The results of the study help clarify the differences in the spatial and temporal distribution of groundwater resources and their driving mechanisms in the northeastern black soil regions and provide a reference for the conservation and sustainable utilization of groundwater resources in the black soil region. Full article

With the rapid development of digital intelligent agriculture, the accurate extraction of field information from remote sensing imagery to guide agricultural planning has become an important issue. In order to better extract fields, we analyze the scale characteristics of agricultural fields and incorporate the multi-scale idea into a Transformer. We subsequently propose an improved deep learning method named the Multi-Swin Mask Transformer (MSMTransformer), which is based on Mask2Former (an end-to-end instance segmentation framework). In order to prove the capability and effectiveness of our method, the iFLYTEK Challenge 2021 Cultivated Land Extraction competition dataset is used and the results are compared with Mask R-CNN, HTC, Mask2Former, etc. The experimental results show that the network has excellent performance, achieving a bbox_AP50 score of 0.749 and a segm_AP50 score of 0.758. Through comparative experiments, it is shown that the MSMTransformer network achieves the optimal values in all the COCO segmentation indexes, and can effectively alleviate the overlapping problem caused by the end-to-end instance segmentation network in dense scenes. Full article

With the wideband beamforming approaches, the synthetic aperture radar (SAR) could achieve high azimuth resolution and wide swath. However, the performance of conventional adaptive wideband time-domain beamforming is severely affected as the received signal snapshots are insufficient for adaptive approaches. In this paper, a wideband beamformer using convolutional neural network (CNN) method, namely, frequency constraint wideband beamforming prediction network (WBPNet), is proposed to obtain a satisfactory performance in the circumstances of scanty snapshots. The proposed WBPNet successfully estimates the direction of arrival of interference with scanty snapshots and obtains the optimal weights with effectively null for the interference by utilizing the uniqueness of CNN to extract potential nonlinear features of input information. Meanwhile, the novel beamformer has an undistorted response to the wideband signal of interest. Compared with the conventional time-domain wideband beamforming algorithm, the proposed method can fast obtain adaptive weights because of using few snapshots. Moreover, the proposed WBPNet has a satisfactory performance on wideband beamforming with low computational complexity because it avoids the inverse operation of covariance matrix. Simulation results show the meliority and feasibility of the proposed approach. Full article

Remote sensing object detection based on the combination of infrared and visible images can effectively adapt to the around-the-clock and changeable illumination conditions. However, most of the existing infrared and visible object detection networks need two backbone networks to extract the features of two modalities, respectively. Compared with the single modality detection network, this greatly increases the amount of calculation, which limits its real-time processing on the vehicle and unmanned aerial vehicle (UAV) platforms. Therefore, this paper proposes a local adaptive illumination-driven input-level fusion module (LAIIFusion). The previous methods for illumination perception only focus on the global illumination, ignoring the local differences. In this regard, we design a new illumination perception submodule, and newly define the value of illumination. With more accurate area selection and label design, the module can more effectively perceive the scene illumination condition. In addition, aiming at the problem of incomplete alignment between infrared and visible images, a submodule is designed for the rapid estimation of slight shifts. The experimental results show that the single modality detection algorithm based on LAIIFusion can ensure a large improvement in accuracy with a small loss of speed. On the DroneVehicle dataset, our module combined with YOLOv5L could achieve the best performance. Full article

Over the last 20 years, there has been a surge of interest in the use of reflectance data collected using satellites and aerial vehicles to monitor vegetation diversity. One methodological option to monitor these systems involves developing empirical relationships between spectral heterogeneity in space (spectral variation) and plant or habitat diversity. This approach is commonly termed the ‘Spectral Variation Hypothesis’. Although increasingly used, it is controversial and can be unreliable in some contexts. Here, we review the literature and apply three-level meta-analytical models to assess the test results of the hypothesis across studies using several moderating variables relating to the botanical and spectral sampling strategies and the types of sites evaluated. We focus on the literature relating to grasslands, which are less well studied compared to forests and are likely to require separate treatments due to their dynamic phenology and the taxonomic complexity of their canopies on a small scale. Across studies, the results suggest an overall positive relationship between spectral variation and species diversity (mean correlation coefficient = 0.36). However, high levels of both within-study and between-study heterogeneity were found. Whether data was collected at the leaf or canopy level had the most impact on the mean effect size, with leaf-level studies displaying a stronger relationship compared to canopy-level studies. We highlight the challenges facing the synthesis of these kinds of experiments, the lack of studies carried out in arid or tropical systems and the need for scalable, multitemporal assessments to resolve the controversy in this field. Full article

Converting fertile, cultivated land into Eucalyptus plantations has become a common practice in Ethiopia. Integrating geospatial techniques with socio-economic data analysis can be a useful method to evaluate the expansion of Eucalyptus and its underlying factors. The objective of this study is to detect the spatio-temporal patterns and main factors contributing to Eucalyptus expansion in the Mecha district of Ethiopia. To quantify the spatial extents of Eucalyptus plantations, the study employed Landsat images from 1991 to 2021 with supervised image classification in ERDAS Imagine 2015. In addition, 120 households were chosen using random sampling technique to incorporate socioeconomic factors related to Eucalyptus expansion. The result shows that, Eucalyptus plantations expanded significantly across the study area during the last three decades. Eucalyptus plantation covered 908.87 ha, 3719.05 ha, and 26261.9 ha in 1991, 2006, and 2021, respectively. The increment was mostly at the expense of fertile cultivated land use. The main reasons for its expansion are linked with farmer’s expectations of a better source of income, apprehension about the detrimental effects on nearby cropland, and its affordable production cost. In conclusion, the study area faces challenges from the uncontrolled expansion of Eucalyptus plantations on productive lands. Therefore, careful management and intervention strategies should be established to manage its rapid expansion. Full article

Sen2-Agri is a software system that was developed to facilitate the use of multi-temporal satellite data for crop classification with a random forest (RF) classifier in an operational setting. It automatically ingests and processes Sentinel-2 and LandSat 8 images. Our goal was to provide practitioners with recommendations for the best sample size and composition. The study area was located in the Yaqui Valley in Mexico. Using polygons of more than 6000 labeled crop fields, we prepared data sets for training, in which the nine crops had an equal or proportional representation, called Equal or Ratio, respectively. Increasing the size of the training set improved the overall accuracy (OA). Gains became marginal once the total number of fields approximated 500 or 40 to 45 fields per crop type. Equal achieved slightly higher OAs than Ratio for a given number of fields. However, recall and F-scores of the individual crops tended to be higher for Ratio than for Equal. The high number of wheat fields in the Ratio scenarios, ranging from 275 to 2128, produced a more accurate classification of wheat than the maximal 80 fields of Equal. This resulted in a higher recall for wheat in the Ratio than in the Equal scenarios, which in turn limited the errors of commission of the non-wheat crops. Thus, a proportional representation of the crops in the training data is preferable and yields better accuracies, even for the minority crops. Full article

Road geometric information and a digital model based on light detection and ranging (LiDAR) can perform accurate geometric inventories and three-dimensional (3D) descriptions for as-built roads and infrastructures. However, unorganized point clouds and complex road scenarios would reduce the accuracy of geometric information extraction and digital modeling. There is a standardization need for information extraction and 3D model construction that integrates point cloud processing and digital modeling. This paper develops a framework from semantic segmentation to geometric information extraction and digital modeling based on LiDAR data. A semantic segmentation network is improved for the purpose of dividing the road surface and infrastructure. The road boundary and centerline are extracted by the alpha-shape and Voronoi diagram methods based on the semantic segmentation results. The road geometric information is obtained by a coordinate transformation matrix and the least square method. Subsequently, adaptive road components are constructed using Revit software. Thereafter, the road route, road entity model, and various infrastructure components are generated by the extracted geometric information through Dynamo and Revit software. Finally, a detailed digital model of the road scenario is developed. The Toronto-3D and Semantic3D datasets are utilized for analysis through training and testing. The overall accuracy (OA) of the proposed net for the two datasets is 95.3 and 95.0%, whereas the IoU of segmented road surfaces is 95.7 and 97.9%. This indicates that the proposed net could accomplish superior performance for semantic segmentation of point clouds. The mean absolute errors between the extracted and manually measured geometric information are marginal. This demonstrates the effectiveness and accuracy of the proposed extraction methods. Thus, the proposed framework could provide a reference for accurate extraction and modeling from LiDAR data. Full article

Although understanding the carbon and water cycles of dryland ecosystems in terms of water use efficiency (WUE) is important, WUE and its driving mechanisms are less understood in Central Asia. This study calculated Central Asian WUE for 2001–2021 based on the Google Earth Engine (GEE) platform and analyzed its spatial and temporal variability using temporal information entropy. The importance of atmospheric factors, hydrological factors, and biological factors in driving WUE in Central Asia was also explored using a geographic detector. The results show the following: (1) the average WUE in Central Asia from 2001–2021 is 2.584–3.607 gCkg⁻¹H₂O, with weak inter-annual variability and significant intra-annual variability and spatial distribution changes; (2) atmospheric and hydrological factors are strong drivers, with land surface temperature (LST) being the strongest driver of WUE, explaining 54.8% of variation; (3) the interaction of the driving factors can enhance the driving effect by more than 60% for the interaction between most atmospheric factors and vegetation factors, of which the effect of the interaction of temperature (TEM) with vegetation cover (FVC) is the greatest, explaining 68.1% of the change in WUE. Furthermore, the interaction of driving factors with very low explanatory power (e.g., water pressure (VAP), aerosol optical depth over land (AOD), and groundwater (GWS)) has a significant enhancement effect. Vegetation is an important link in driving WUE, and it is important to understand the mechanisms of WUE change to guide ecological restoration projects. Full article

Mangrove wetlands are rapidly being lost due to anthropogenic disturbances and natural processes, such as sea-level rise (SLR), but are also recovering as a result of conservation efforts. Accurate and contemporary mangrove maps to detect their distribution and changes are urgently needed to understand how mangroves respond to global change and develop effective conservation projects. Here, we developed a new change detection algorithm called temporal consistency checking combining annual classification and spectral time series (TCC-CS) for tracking mangrove losses and gains. Specifically, mangrove change events were determined by measuring the deviation of greenness and wetness of candidate change segments from automatically collected mangrove reference samples. By applying to the world’s largest mangrove patches, we monitored the 35-year mangrove trajectory in the Sundarbans from 1988 to 2022 using all available Landsat images on the Google Earth Engine platform. In the Sundarbans, 18,501.89 ha of mangroves have been gained, but these have been offset by losses of 27,009.79 ha, leading to a net mangrove loss of 1.42% (8507.9 ha) in the past 35 years. We further mapped the pixel-level change agents and found that SLR-induced erosion and degradation, instead of human activities, were the major drivers of losses in the Sundarbans. Trend analysis on loss agents indicates that mangrove losses caused by human activities, such as the expansion of croplands and aquaculture ponds, have declined, but SLR is still a persistent threat to mangrove wetlands in this iconic mangrove area. Our study provides a computationally efficient methodology for examining large-scale mangrove changes, and the resultant annual mangrove maps provide strong support for mangrove conservation in the Sundarbans. Full article

Cloud occlusion phenomena are widespread in optical remote sensing (RS) images, leading to information loss and image degradation and causing difficulties in subsequent applications such as land surface classification, object detection, and land change monitoring. Therefore, thin cloud removal is a key preprocessing procedure for optical RS images, and has great practical value. Recent deep learning-based thin cloud removal methods have achieved excellent results. However, these methods have a common problem in that they cannot obtain large receptive fields while preserving image detail. In this paper, we propose a novel wavelet-integrated convolutional neural network for thin cloud removal (WaveCNN-CR) in RS images that can obtain larger receptive fields without any information loss. WaveCNN-CR generates cloud-free images in an end-to-end manner based on an encoder–decoder-like architecture. In the encoding stage, WaveCNN-CR first extracts multi-scale and multi-frequency components via wavelet transform, then further performs feature extraction for each high-frequency component at different scales by multiple enhanced feature extraction modules (EFEM) separately. In the decoding stage, WaveCNN-CR recursively concatenates the processed low-frequency and high-frequency components at each scale, feeds them into EFEMs for feature extraction, then reconstructs the high-resolution low-frequency component by inverse wavelet transform. In addition, the designed EFEM consisting of an attentive residual block (ARB) and gated residual block (GRB) is used to emphasize the more informative features. ARB and GRB enhance features from the perspective of global and local context, respectively. Extensive experiments on the T-CLOUD, RICE1, and WHUS2-CR datasets demonstrate that our WaveCNN-CR significantly outperforms existing state-of-the-art methods. Full article

Negative air ions (NAIs), which are known as the “air vitamin”, have been widely used as a measure of air cleanness. Field observation provides an alternative way to record site-level NAIs. However, these observations fail to capture the regional distribution of NAIs due to the limited number of sites. In this study, satellite-based bio-geophysical parameters from the climate, topography, air quality, vegetation, and anthropogenic intensity were used to estimate the daily NAIs with the Random Forest model (RF). In situ NAI observations over Zhejiang Province, China were incorporated into the model. Daily NAIs were averaged to capture the spatio-temporal distribution. The results showed that (1) the RF algorithm performed better than traditional regression analysis and the common BP neural network to generate regional NAIs at a spatial scale of 500 m over the larger scale, with an RMSE of 258.62, R² of 0.878 for model training, and R² of 0.732 for model testing; (2) in the variable importance measures (VIM) analysis, 87.96% of the NAI variance was caused by the elevation, aspect, slope, surface temperature, solar-induced chlorophyll fluorescence (SIF), relative humidity (RH), and the concentration of carbon monoxide (CO), while path analysis indicated that SIF was one of the most important factors affecting NAI concentration across the whole region; (3) NAI concentrations in 87.16% of the region were classified above grade III (>500 ions cm⁻³), which was able to meet the needs of human health maintenance; (4) the highest NAI concentration was distributed over the southwest of the Zhejiang Province, where forest land dominates. The lowest NAI concentration was mostly found in the northeast regions, where urban areas are well-developed; and (5) among different land types, the NAI concentrations were ranked as forest land > water bodies > barren > grassland > croplands > urban and built-up. Among different seasons, summer and winter have the highest and lowest NAIs, respectively. Our study provided a substantial reference for ecosystem services assessment in Zhejiang Province. Full article

Hyperspectral data usually consists of hundreds of narrow spectral bands and provides more detailed spectral characteristics compared to commonly used multispectral data in remote sensing applications. However, highly correlated spectral bands in hyperspectral data lead to computational complexity, which limits many applications or traditional methods when applied to hyperspectral data. The dimensionality reduction of hyperspectral data becomes one of the most important pre-processing steps in hyperspectral data analysis. Recently, deep reinforcement learning (DRL) has been introduced to hyperspectral data band selection (BS); however, the current DRL methods for hyperspectral data BS simply remove redundant bands, lack the significance analysis for the selected bands, and the reward mechanisms used in DRL only take basic forms in general. In this paper, a new reward mechanism strategy has been proposed, and Double Deep Q-Network (DDQN) is introduced during BS using DRL to improve the network stabilities and avoid local optimum. To verify the effect of the proposed BS method, land cover classification experiments were designed and carried out to analyze and compare the proposed method with other BS methods. In the land cover classification experiments, the overall accuracy (OA) of the proposed method can reach 98.37%, the average accuracy (AA) is 95.63%, the kappa coefficient (Kappa) is 97.87%. Overall, the proposed method is superior to other BS methods. Experiments have also shown that the proposed method works not only for airborne hyperspectral data (AVIRIS and HYDICE), but also for hyperspectral satellite data, such as PRISMA data. When hyperspectral data is applied to similar applications, the proposed BS method could be a candidate for the BS preprocessing options. Full article

This paper presents a novel two-view Structure-from-Motion (SfM) algorithm with the application of multiple Feature Detector Operators (FDO). The key of this study is the implementation of multiple FDOs into a two-view SfM algorithm. The two-view SfM algorithm workflow can be divided into three general steps: feature detection and matching, pose estimation and point cloud (PCL) generation. The experimental results, the quantitative analyses and a comparison with existing algorithms demonstrate that the implementation of multiple FDOs can effectively improve the performance of a two-view SfM algorithm. Firstly, in the Oxford test dataset, the RMSE reaches on average 0.11 m (UBC), 0.36 m (bikes), 0.52 m (trees) and 0.37 m (Leuven). This proves that illumination changes, blurring and JPEG compression can be handled satisfactorily. Secondly, in the EPFL dataset, the number of features lost in the processes is 21% with a total PCL of 27,673 pt, and this is only minimally higher than ORB (20.91%) with a PCL of 10,266 pt. Finally, the verification process with a real-world unmanned aerial vehicle (UAV) shows that the point cloud is denser around the edges, the corners and the target, and the process speed is much faster than existing algorithms. Overall, the framework proposed in this study has been proven a viable alternative to a classical procedure, in terms of performance, efficiency and simplicity. Full article