Very High Resolution Research Articles

Multiscale Semantic Guidance Network for Object Detection in VHR Remote Sensing Images

With the development of convolutional neural network (CNN), many CNN-based object detection methods have made a remarkable success in very high-resolution (VHR) remote sensing images (RSIs). However, the standard convolution has a fixed receptive field, which makes it deficient in dynamic feature capture; complex backgrounds may also lead to the degradation of detection performance. Accordingly, this letter proposes a novel multiscale semantic guidance network (MSGN) to tackle these problems, wherein, based on the deformable convolution, an improved feature extraction backbone is proposed to capture features dynamically. Moreover, features from different layers are used to ensure the ability for detecting multiscale objects. Furthermore, a multilevel semantic guidance filtering subnetwork is proposed based on the designed backward semantic guidance filtering (BSGF) module, to suppress the complex backgrounds. Experimental results show that the proposed MSGN has stronger robustness and a better accuracy for multiscale object detection, compared with other reference methods.

Cascaded Multi-Task Road Extraction Network for Road Surface, Centerline, and Edge Extraction

Road extraction from very high-resolution (VHR) remote sensing imagery remains a huge challenge, due to the shadows and occlusions of trees and buildings. Such complex backgrounds result in deep networks often producing fragmented roads with poor connectivity. Road extraction has three typical tasks: road surface segmentation (SS), centerline extraction (CE), and edge detection (ED), which are conducted in a wide range of real applications. Also, the three tasks have a symbiotic relationship, i.e., the road SS determines the location of the centerline and edges, and the CE and ED can allow the generation of more continuous road surfaces. However, most of the previous works have completed these three tasks separately, without exploiting the symbiotic relationship between them to boost the road connectivity. In this article, in order to improve road connectivity, a cascaded multitask (CasMT) road extraction framework for simultaneously extracting the road surface, centerline, and edges is proposed. In the proposed framework, topology-aware learning is applied to capture the long-distance topological relationships, and hard example mining (HEM) loss is employed to focus more on hard samples, to further enhance the road completeness. Extensive experiments were conducted on the DeepGlobe road dataset and a large-scale road dataset (called the LSCC dataset) from the three Chinese cities of Beijing, Shanghai, and Wuhan. The experimental results obtained on the public DeepGlobe dataset demonstrate that the proposed CasMT framework can significantly outperform the current state-of-the-art method. Moreover, the generalization capability of the model was verified on the LSCC dataset, where the proposed CasMT framework achieved the best performance in the average path length similarity (APLS) road topology metric, which further confirms the superiority of the proposed framework.

Change Detection in VHR Imagery With Severe Co-Registration Errors Using Deep Learning: A Comparative Study

Change detection (CD) through Earth observation techniques can offer very significant information for monitoring tasks in a time-efficient manner. Very high-resolution (VHR) images can display objects in fine detail, thus making it possible to rapidly perceive isolated changes. However, this is a challenging task because of the increased within-class variance and geometric registration errors caused by different satellite view directions and angles. Lately, deep learning (DL) CD methods have proven very appealing for the CD problem because of their flexibility to combine and process different types of information along with the increased availability of higher processing power systems. Even though previous research has developed several notable DL methodologies, it has mostly focused on images with minor co-registration errors. Based on that, the goal of this study is to evaluate the performance of five state-of-the-art DL CD methods, two unsupervised and three supervised, on VHR images with severe co-registration errors. The methods are implemented on four urban European areas of versatile morphology. In addition, before applying the CD process, four popular automatic co-registration methods were evaluated because of the importance of this pre-processing step for the successful output of the CD problem. It was shown that phase correlation used on the Fourier-Mellin Transform produced the most satisfactory co-registration results and STANet detected building-related changes most successfully. Its success can be attributed to its particular attention mechanism and its training dataset. The rest of the co-registration and CD methods showed low performance.

Object-Based Sorted-Histogram Similarity Measurement for Detecting Land Cover Change With VHR Remote Sensing Images

Land cover change detection (LCCD) with very high-resolution (VHR) remote sensing images has been widely used in various applications. However, pseudo-changes and noise usually affect the performance of detection map. In this letter, an object-oriented sorted-histogram similarity measurement (OSSM) is proposed for measuring the change magnitude between bi-temporal remote sensing images. First, multi-scale objects are acquired for the post-event image using a multi-scale segmentation algorithm, and then the pixels within each object are considered to construct the pairwise histograms and the bin of each histogram is sorted in descending order. Second, a bin-to-bin (B2B) distance is defined to measure the change magnitude between the pairwise object-based histograms, and the change magnitude image (CMI) is generated after all the bi-temporal images are scanned object by object. Finally, a simple yet effective method called Otsu is used to divide the CMI into binary change detection maps. The experiments on three pairs of VHR images produced promising results compared with five popular LCCD approaches, for example, the improvement is about 2.5% for F-score.

Consistency-Regularized Region-Growing Network for Semantic Segmentation of Urban Scenes With Point-Level Annotations.

Deep learning algorithms have obtained great success in semantic segmentation of very high-resolution (VHR) remote sensing images. Nevertheless, training these models generally requires a large amount of accurate pixel-wise annotations, which is very laborious and time-consuming to collect. To reduce the annotation burden, this paper proposes a consistency-regularized region-growing network (CRGNet) to achieve semantic segmentation of VHR remote sensing images with point-level annotations. The key idea of CRGNet is to iteratively select unlabeled pixels with high confidence to expand the annotated area from the original sparse points. However, since there may exist some errors and noises in the expanded annotations, directly learning from them may mislead the training of the network. To this end, we further propose the consistency regularization strategy, where a base classifier and an expanded classifier are employed. Specifically, the base classifier is supervised by the original sparse annotations, while the expanded classifier aims to learn from the expanded annotations generated by the base classifier with the region-growing mechanism. The consistency regularization is thereby achieved by minimizing the discrepancy between the predictions from both the base and the expanded classifiers. We find such a simple regularization strategy is yet very useful to control the quality of the region-growing mechanism. Extensive experiments on two benchmark datasets demonstrate that the proposed CRGNet significantly outperforms the existing state-of-the-art methods. Codes and pre-trained models are available online (https://github.com/YonghaoXu/CRGNet).

Gated Recurrent Multiattention Network for VHR Remote Sensing Image Classification

With the advances of deep learning, many recent CNN-based methods have yielded promising results for image classification. In very high-resolution (VHR) remote sensing images, the contributions of different regions to image classification can vary significantly, because informative areas are generally limited and scattered throughout the whole image. Therefore, how to pay more attention to these informative areas and better incorporate them over long distances are two main challenges to be addressed. In this article, we propose a gated recurrent multiattention neural network (GRMA-Net) to address these problems. Because informative features generally occur at multiple stages in a network (i.e., local texture features at shallow layers and global profile features at deep layers), we use multilevel attention modules to focus on informative regions to extract more discriminative features. Then, these features are arranged as spatial sequences and fed into a deep-gated recurrent unit (GRU) to capture long-range dependency and contextual relationship. We evaluate our method on the UC Merced (UCM), Aerial Image dataset (AID), NWPU-RESISC (NWPU), and Optimal-31 (Optimal) datasets. Experimental results have demonstrated the superior performance of our method as compared to other state-of-the-art methods.

Multitask Semantic Boundary Awareness Network for Remote Sensing Image Segmentation

In remote sensing images, boundary information plays a crucial role in land-cover segmentation. However, it is a challenging problem that sufficiently extracts complete and sharp boundaries from complex very-high-resolution (VHR) remote sensing images. To tackle this problem, we propose a semantic boundary awareness network (SBANet). The SBANet captures refined boundary information of land covers in feature extraction and then supervises its learning with a designed boundary loss. The key of SBANet includes boundary attention module (BA-module) and adaptive weights of multitask learning (AWML). The BA-module is proposed to capture land-cover boundary information from hierarchical features aggregation in a bottom-up manner. It emphasizes useful boundary information and relieves noise information in low-level features with the guidance of high-level features. To directly learn the boundary information, AWML adds a boundary loss to the original semantic loss by an adaptive fusion manner. This multitask learning enables the semantic information and the boundary information to work collaboratively and promote each other. Note that the BA-module and AWML are plug-and-play. Experimental results demonstrate the effectiveness of the proposed SBANet on the available ISPRS 2-D semantic labeling Potsdam and Vaihingen data sets. The SBANet also achieves the state-of-the-art performance in terms of overall accuracy (OA) and mean <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$F_{1}$ </tex-math></inline-formula> score (m- <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$F_{1}$ </tex-math></inline-formula> ).

Difference Enhancement and Spatial–Spectral Nonlocal Network for Change Detection in VHR Remote Sensing Images

The popular Siamese convolutional neural networks (CNNs) for remote sensing (RS) image change detection (CD) often suffer from two problems. First, they either ignore the original information of bitemporal images or insufficiently utilize the difference information between bitemporal images, which leads to the low tightness of the changed objects. Second, Siamese CNNs always employ dual-branch encoders for CD, which increases computational cost. To address the above issues, this article proposes a network based on difference enhancement and spatial&#x2013;spectral nonlocal (DESSN) for CD in very-high-resolution (VHR) images. This article makes threefold contributions. First, we design a difference enhancement (DE) module that can effectively learn the difference representation between foreground and background to reduce the impact of irrelevant changes on the detection results. Second, we present a spatial&#x2013;spectral nonlocal (SSN) module that is different from vanilla nonlocal because multiscale spatial global features are incorporated to model the large-scale variation of objects during CD. The module can be used to strengthen the edge integrity and internal tightness of changed objects. Third, the asymmetric double convolution with Ghost (ADCG) module is exploited instead of standard convolution. The ADCG can not only refine the edge information of the changed objects, since horizontal and vertical convolutional kernels have good contour preservation advantages, but also greatly reduce the computational complexity of the proposed model. The experiments on two public VHR CD datasets demonstrate that the proposed network can provide higher detection accuracy and requires smaller memory usage than state-of-the-art networks.

Cascaded Multiscale Structure With Self-Smoothing Atrous Convolution for Semantic Segmentation

Convolutional neural networks (CNNs) have attracted great attention in the semantic segmentation of very-high-resolution (VHR) images of urban areas. However, large-scale variation of objects in the urban areas often makes it difficult to achieve good segmentation accuracy. Atrous convolution and atrous spatial pyramid pooling composed of atrous convolution can alleviate this problem by exploring multiscale contextual information. Unfortunately, atrous convolution causes gridding artifacts, where actual receptive fields are separated unit sets and fail to cover all the receptive fields. To address this problem, in this article, we first propose a self-smoothing atrous convolution (SS-AConv) that intrinsically improves atrous convolution, unlike existing methods. SS-AConv enhances sampling rates with low computational costs by adding several key parameters in the spatial dimension of its filter. Then, in the backbone, the Xception network, all max pooling operations are replaced with SS-AConvs for a large and effective receptive field. Moreover, to extract diverse features at multiple scales, we propose the SS-AConv cascaded multiscale structure (SCMS) by integrating SS-AConvs with different rates and the residual correction scheme (RCS) into a cascaded spatial pyramid. Finally, to extract diverse features at dense multiple scales, the SS-AConv convolutional network (SS-ACNet) is constructed by integrating SCMS into both the encoder and decoder layers of the modified Xception network. Extensive experimental results show that SS-ACNet outperforms some state-of-the-art methods on four open challenge datasets: the ISPRS Vaihingen and Potsdam, Cityscapes, and PASCAL VOC 2012.

Object-Based Classification Framework of Remote Sensing Images With Graph Convolutional Networks

Object-based image classification (OBIC) on very-high-resolution (VHR) remote sensing (RS) images is utilized in a wide range of applications. Nowadays, many existing OBIC methods only focus on features of each object itself, neglecting the contextual information among adjacent objects and resulting in low classification accuracy. Inspired by a spectral graph theory, we construct a graph structure from objects generated from VHR RS images and propose an OBIC framework based on truncated sparse singular value decomposition and graph convolutional network (GCN) model, aiming to make full use of relativities among objects and produce an accurate classification. Through conducting experiments on two annotated RS image data sets, our framework obtained 97.2% and 66.9% overall accuracy, respectively, in automatic and manual object segmentation circumstances, within a processing time of about 1/100 of convolutional neural network (CNN)-based methods’ training time.

Semantic Segmentation of Remote-Sensing Images Through Fully Convolutional Neural Networks and Hierarchical Probabilistic Graphical Models

Deep learning (DL) is currently the dominant approach to image classification and segmentation, but the performances of DL methods are remarkably influenced by the quantity and quality of the ground truth (GT) used for training. In this article, a DL method is presented to deal with the semantic segmentation of very-high-resolution (VHR) remote-sensing data in the case of scarce GT. The main idea is to combine a specific type of deep convolutional neural networks (CNNs), namely fully convolutional networks (FCNs), with probabilistic graphical models (PGMs). Our method takes advantage of the intrinsic multiscale behavior of FCNs to deal with multiscale data representations and to connect them to a hierarchical Markov model (e.g., making use of a quadtree). As a consequence, the spatial information present in the data is better exploited, allowing a reduced sensitivity to GT incompleteness to be obtained. The marginal posterior mode (MPM) criterion is used for inference in the proposed framework. To assess the capabilities of the proposed method, the experimental validation is conducted with the ISPRS 2D Semantic Labeling Challenge datasets on the cities of Vaihingen and Potsdam, with some modifications to simulate the spatially sparse GTs that are common in real remote-sensing applications. The results are quite significant, as the proposed approach exhibits a higher producer accuracy than the standard FCNs considered and especially mitigates the impact of scarce GTs on minority classes and small spatial details.

A Semisupervised CRF Model for CNN-Based Semantic Segmentation With Sparse Ground Truth

Convolutional neural networks (CNNs) represent the new reference approach for semantic segmentation of very-high-resolution (VHR) images, due to their ability to automatically capture semantic information while learning relevant features. However, as for most supervised methods, the map accuracy depends on the quantity and quality of ground truth (GT) used to train them. The use of densely annotated data (i.e., a detailed, exhaustive, pixel-level GT) allows to obtain effective CNN models but normally implies high efforts in annotation. Such ground truth is often available in benchmark datasets on which new methods are tested, but not on real data for land-cover applications, where only sparse annotations might be sufficiently cost effective. A CNN model trained with such incomplete GT maps has the tendency to smooth object boundaries because they are never precisely delineated in the GT. To cope with those shortcomings, we propose to exploit the intermediate activation maps of the CNN and to deploy a semisupervised fully connected conditional random field (CRF). In comparison with competitors using the same sparse annotations, the proposed method is able to better fill part of the performance gap compared to a CNN trained on the densely annotated, but generally unavailable, GTs.

Multilevel Deformable Attention-Aggregated Networks for Change Detection in Bitemporal Remote Sensing Imagery

Deep learning (DL) approaches based on convolutional encoder&#x2013;decoder networks have shown promising results in bitemporal change detection. However, their performance is limited by insufficient contextual information aggregation because they cannot fully capture the implicit contextual dependency relationships among feature maps at different levels. Moreover, harvesting long-range contextual information typically incurs high computational complexity. To circumvent these challenges, we propose multilevel deformable attention-aggregated networks (MLDANets) to effectively learn long-range dependencies across multiple levels of bitemporal convolutional features for multiscale context aggregation. Specifically, a multilevel change-aware deformable attention (MCDA) module consisting of linear projections with learnable parameters is built based on multihead self-attention (SA) with a deformable sampling strategy. It is applied in the skip connections of an encoder&#x2013;decoder network taking a bitemporal deep feature hypersequence (BDFH) as input. MCDA can progressively address a set of informative sampling locations in multilevel feature maps for each query element in the BDFH. Simultaneously, MCDA learns to characterize beneficial information from different spatial and feature subspaces of BDFH using multiple attention heads for change perception. As a result, contextual dependencies across multiple levels of bitemporal feature maps can be adaptively aggregated via attention weights to generate multilevel discriminative change-aware representations. Experiments on very-high-resolution (VHR) datasets verify that MLDANets outperform state-of-the-art change detection approaches with dramatically faster training convergence and high computational efficiency.

Allometric Equations to Estimate the Dry Mass of Sahel Woody Plants from Very-High Resolution Satellite Imagery

Allometric Equations to Estimate the Dry Mass of Sahel Woody Plants from Very-High Resolution Satellite Imagery

FPGA-Based Implementation of a CNN Architecture for the On-Board Processing of Very High-Resolution Remote Sensing Images

Over the last years, Convolutional Neural Networks (CNNs) have been widely used in remote sensing applications, such as marine surveillance, traffic management or road networks detection. However, since CNNs have extremely high computation, bandwith and memory requirements, the hardware implementation of a CNN on space-grade devices like FPGAs for the on-board processing of the acquired images has brought many challenges. In fact, FPGAs show great advantages in terms of reconfigurability and performance-power ratio, whichmakes them a suitable election to deploy CNNs, although the computational capabilities of this type of hardware devices available on-board is limited, so implementations have to be carefully planned. In this paper, the authors present their work towards the implementation of an efficient CNN onto a spacegrade FPGA in order to achieve the on-board processing of very-high resolution remotely sensed images as soon as the data are provided by the sensor. All this work has been conducted within the EU-funded VIDEO project. As it will be presented in this paper, the work includes the introduction of a methodology based on the project constraints, the evaluation of different state-of-the-art CNN architectures by means of a new efficiency measurement also proposed in this work, the introduction of a new efficient CNN architecture, and finally, its optimized hardware implementation by means of high-level synthesis tools.

CMGFNet: A deep cross-modal gated fusion network for building extraction from very high-resolution remote sensing images

The extraction of urban structures such as buildings from very high-resolution (VHR) remote sensing imagery has improved dramatically, thanks to recent developments in deep multimodal fusion models. However, Due to the variety of colour intensities with complex textures of building objects in VHR images and the low quality of the digital surface model (DSM), it is challenging to develop the optimal cross-modal fusion network that takes advantage of these two modalities. This research presents an end-to-end cross-modal gated fusion network (CMGFNet) for extracting building footprints from VHR remote sensing images and DSMs data. The CMGFNet extracts multi-level features from RGB and DSM data by using two separate encoders. We offer two methods for fusing features in two modalities: Cross-modal and multi-level feature fusion. For cross-modal feature fusion, a gated fusion module (GFM) is proposed to combine two modalities efficiently. The multi-level feature fusion fuses the high-level features from deep layers with shallower low-level features through a top-down strategy. Furthermore, a residual-like depth-wise separable convolution (R-DSC) is introduced to enhance the performance of the up-sampling process and decrease the parameters and time complexity in the decoder section. Experimental results from challenging datasets show that the CMGFNet outperforms other state-of-the-art models. The efficacy of all significant elements is also confirmed by the extensive ablation study.

Very High-Resolution Satellite-Derived Bathymetry and Habitat Mapping Using Pleiades-1 and ICESat-2

Accurate and reliable bathymetric data are needed for a wide diversity of marine research and management applications. Satellite-derived bathymetry represents a time saving method to map large shallow waters of remote regions compared to the current costly in situ measurement techniques. This study aims to create very high-resolution (VHR) bathymetry and habitat mapping in Mayotte island waters (Indian Ocean) by fusing 0.5 m Pleiades-1 passive multispectral imagery and active ICESat-2 LiDAR bathymetry. ICESat-2 georeferenced photons were filtered to remove noise and corrected for water column refraction. The bathymetric point clouds were validated using the French naval hydrographic and oceanographic service Litto3D® dataset and then used to calibrate the multispectral image to produce a digital depth model (DDM). The latter enabled the creation of a digital albedo model used to classify benthic habitats. ICESat-2 provided bathymetry down to 15 m depth with a vertical accuracy of bathymetry estimates reaching 0.89 m. The benthic habitats map produced using the maximum likelihood supervised classification provided an overall accuracy of 96.62%. This study successfully produced a VHR DDM solely from satellite data. Digital models of higher accuracy were further discussed in the light of the recent and near-future launch of higher spectral and spatial resolution satellites.

Mapping standing dead trees in temperate montane forests using a pixel- and object-based image fusion method and stereo WorldView-3 imagery

Information about the distribution of standing dead trees (SDT) is essential for forest biodiversity estimation, forest disturbances monitoring, and forest management strategy planning. Although remote sensing techniques offer unique capabilities to map SDT over large areas, three major hurdles exist: (1) the sporadic distribution of SDT in the study area; (2) often poor spectral separability between SDT and bare ground in forests; (3) the prominent spectral variability within SDT due to variations in background effect and canopy illumination. To address these problems, we proposed a pixel- and object-based image fusion (POBIF) approach using very high-resolution (VHR) stereo WorldView-3 (WV-3) data. The stereo WV-3 derived spectral bands, canopy height model (CHM), vegetation indices (VIs), and texture features were used as inputs in six classification scenarios with different variable combinations. A deep learning algorithm, deep neural network (DNN), and two machine learning algorithms, support vector machine (SVM) and random forest (RF), were utilized to process the pixel-based (PB) and object-based (OB) information. All PB and OB classifiers were then combined using a stacked generalization strategy to develop the POBIF model. Comparing the six scenarios we assessed the importance of the CHM, VIs, and textures for accurate SDT mapping. As a result, we found (1) the POBIF outperformed both PB and OB methods for SDT mapping, generating notably higher F1-score (p < 0.05); (2) Adding the CHM reduced the commission and omission errors caused by bare ground and artificial surfaces, and the highest classification accuracy was achieved when combing all the WV-3 derived variables; (3) SDT covered 0.89% of the forest areas in the study area and was particularly distributed on higher and steeper north-facing (northeast and northwest) slopes. Large tracts of SDT were found in the strictly protected forests. The study highlights the potential of VHR stereo WV-3 imagery and the POBIF for SDT mapping in a temperate montane forest area with high accuracy. The map created in this study could be used for guiding field investigations and for planning management measures.

Deformation monitoring and thematic mapping of the Badaling Great Wall using very high-resolution interferometric synthetic aperture radar data

The preventive monitoring and sustainable conservation of large-scale cultural heritage sites require satellite-based Earth observations. In this study, we present the first monitoring and thematic mapping results of the Badaling Great Wall (Beijing, China) using deformation indicators calculated from very high-resolution (VHR) multi-temporal spotlight TerraSAR-X data. The proposed coarse–fine search algorithm achieved high computational efficiency for calculating the unknown parameters of the two-tier network persistent scatterer synthetic aperture radar (SAR) interferometry (PSInSAR) approach. The spatiotemporal deformation anomalies, characterized by the absolute velocity, deformation deviation and acceleration, are informative to identify suspected hotspots for prioritizing monitoring activities. We provide an understandable method for thematic mapping and subsequent sustainable conservation of heritage sites by synergistically exploiting impacts from natural degradation and the tourism industry. We determine that the optimum tourist capacity of the site could be 1.0 million per month based on comparing pre- and post-COVID data (2019–2020). This study demonstrates the potential and performance of spaceborne PSInSAR tools for the intelligent management of large-scale architectural heritage sites by integrating InSAR deformation products with environmental and social data.

Mapping fractional woody cover in an extensive semi-arid woodland area at different spatial grains with Sentinel-2 and very high-resolution data

Woody canopy cover is an essential variable to characterize and monitor vegetation health, carbon accumulation and land–atmosphere exchange processes. Remote sensing-based global woody and forest cover maps are available, yet with varying qualities. In arid and semi-arid areas, existing global products often underestimate the presence of woody cover due to the sparse woody cover and bright soil background. Case studies on smaller regions have shown that a combination of collected field data and medium-to-high resolution free satellite data (e.g., Landsat / Sentinel-2) can provide woody cover estimates with practically-sufficient accuracies. However, most earlier studies focused on comparably small regions and relied on costly field data. Here, we present a fully remote sensing-based work-flow to derive woody cover estimates over an area covering more than 0.5 million km2. The work-flow is showcased over the Zagros Mountains, a semi-arid mountain range covering western Iran, the northeast of Iraq and some smaller fraction of southeast Turkey. We use the Google Earth Engine to create homogeneous Sentinel-2 mosaics of the region using data from several years. These data are combined with reference woody cover values derived by a semi-automatic procedure from Google® and Bing® very high resolution (VHR) imagery. Several random forest (RF) models at different spatial grains were trained and at each grain validated with iterative splits of the reference data into training and validation sets (100 repetitions). Best results (considering the trade-off between model performance and spatial detail) were obtained for the model with 40 m spatial grain which showed stable relationships between the VHR-derived reference data and the Sentinel-2 based estimates of woody cover density. The model resulted in median values of coefficient of determination (R2) and RMSE of 0.67 and 0.11, respectively. Our work-flow is potentially also applicable to other arid and semi-arid regions and can contribute to improve currently available global woody cover products, which often perform poorly in semi-arid and arid regions. Comparisons between our woody cover products with common global woody or forest-cover products indicate a clear superiority of our approach. In future studies, these results may be further improved by taking into account regional differences in the drivers of woody-cover patterns along the environmental gradient of the Zagros area.