ABSTRACT The interpretation of high‐resolution remote sensing imagery is essential for accurate land use classification. Optical and synthetic aperture radar (SAR) imagery exhibit complementary characteristics. Their fusion offers an effective approach to mitigating the limitations of single‐modal data and improving classification performance. However, the modal heterogeneity and complexity of optical and SAR imagery pose significant challenges for effective fusion. To address these issues, we propose a heterogeneous adaptive fusion network (HAFNet). First, the multi‐modality feature extractor leverages HRNet to retain the spatial details and local textures of optical images, while a lightweight SparseDense Transformer captures the global structural patterns of SAR data. Within M2FE, the multi‐branch integrated feature enhancement further strengthens the extracted features by emphasizing essential semantic information and suppressing noise. Second, the adaptive multi‐scale attention fusion module employs multi‐scale channel and spatial attention to capture critical information from both modalities, and incorporates a gating mechanism to adjust fusion weights, thereby dynamically exploiting cross‐modal complementarity. Finally, the U‐shaped framework with skip connections enhanced by a dynamic channel fusion module restores spatial resolution and improves the recognition accuracy of small‐scale land cover categories. To validate HAFNet, we conducted extensive comparisons with state‐of‐the‐art models on three public datasets with different resolutions. Experimental results demonstrate that HAFNet achieves improvements of approximately 1.2% in overall accuracy, 1.3% in the Kappa coefficient, 1.4% in F1‐score, and 1.6% in mean Intersection over Union, confirming its effectiveness in land use classification.

ABSTRACTThe rapid development of deep learning techniques has revolutionized various remote sensing applications, which is especially true for change detection (CD) areas. Consequently, the past few years have seen a surge of deep learning change detection (DLCD) techniques with unparalleled improvements in precision, efficiency, and automation. Despite their huge success, these methods often follow a data‐driven routine, where massive labeled data is required to guarantee network parameter learning. However, it is costly and labor‐intensive to obtain sufficient labeled data for the CD task, especially pixel‐level annotations. In this context, label‐efficient DLCD (LE‐DLCD) techniques have garnered increasing attention, which are capable of training CD networks with incomplete labels, inexact labels, or even no explicit labels. In this review, we conducted a comprehensive survey of state‐of‐the‐art label‐efficient DLCD methods, which are categorized into six schemes of semi‐supervised CD, weakly supervised CD, self‐supervised CD, active learning CD, few‐shot CD, and unsupervised CD. Subsequently, each scheme is further categorized into finer subcategories for in‐depth summarization and analysis. Next, we make systematic quantitative comparisons of typical LE–DLCD methods to provide valuable guidance for real‐world scenarios. Finally, the challenges and future directions of LE‐DLCD are presented in detail, which aims to shed light and inspiration on this area for the CD community. To facilitate transparency, we have shared the selected LE‐DLCD methods via https://github.com/daifeng2016/Awesome‐Label‐efficient‐Deep‐Learning‐Change‐Detection‐Methods.

ABSTRACTVisual–Linguistic large models (V‐LLMs) have demonstrated strong semantic understanding capabilities in understanding visual open‐world scenes. However, there is limited research on how to effectively integrate these models with traditional deep learning networks for classify and segment 3D point clouds. This paper explores the joint training mechanisms of V‐VLLMs and Convolutional Neural Networks (CNNs), constructing a pluggable module that enhances 3D object classification and component segmentation using 2D images, 3D point clouds, and language descriptions. The idea of this pluggable module is to employ contrastive learning between the generalized features extracted by large models from text and image features extracted by the CNN to enhance the robustness and understanding accuracy of the original CNN‐based point cloud network. For experiments, we validated the effectiveness and accuracy of the proposed pluggable module using multiple point cloud classification frameworks, based on datasets such as ModelNet40, ScanObjectNN, and ShapeNetCore. The experimental results demonstrate that the multimodal contrastive learning pluggable module significantly enhances performance across various deep learning frameworks, resulting in significant enhancements in classification and segmentation precision. Additionally, for the multiscale feature PointNet++ framework, the proposed multiscale contrastive learning mechanism provides further performance gains. The Code is available at https://github.com/winbigqian/MCL‐3D‐VCNN.

ABSTRACTReliable feature point matching is a pivotal step in the integrated 3D reconstruction of aerial and ground images. Nonetheless, the substantial viewpoint and scale differences between aerial and ground images introduce significant challenges in feature point matching. To address this issue, this study presents a robust matching method for feature points from these two types of images, employing partitioned attention aggregation along with a coarse‐to‐fine hierarchical matching approach, enabling efficient matching of large‐scale aerial‐ground images even with hardware limitations. In the coarse matching phase, region segmentation is carried out on the ground image based on its depth map, which is computed using a monocular depth estimation network. Subsequently, a graph neural network serves as an intermediary for information exchange between features, and partitioned attention aggregation is performed while adhering to the constraints imposed by the segmented image regions. During this process, feature points are grouped to form distinct subgraphs, enhancing the structural similarity between corresponding graphs when there are significant depth and viewpoint changes. This also reduces redundant connections within graph neural networks, leading to efficient feature matching. Corresponding local patches are generated from the full‐sized aerial‐ground images using the initial matches, and further feature point extraction and matching from these patches yield accurate point matches. The experimental results indicate the proposed method performs better than state‐of‐the‐art matching methods. Specifically, compared with the best‐performing method among the comparative ones, our method enhances the average number of correct matches and the average matching precision by approximately 6‐fold and 36%, respectively.

ABSTRACTLeaf chlorophyll content (LCC) is an important indicator for vegetation status monitoring. Remote sensing data offers a nondestructive method for quick estimation of LCC. However, the accuracy of LCC estimation based on remote sensing data depends mainly on the band information. Therefore, it is important to effectively evaluate the effect of bandwidth and position on LCC estimation to select the optimal bandwidth and position. This paper analyzes the effects of different bandwidths and positions on the accuracy of LCC inversion based on single‐band or classical two‐band vegetation indices. First, simulated spectral datasets with different bandwidths and positions are generated by the PROSAIL model, and five classical two‐band vegetation indices are constructed. Then, the sensitivity of single‐band and vegetation indices to changes in bandwidth was evaluated by analyzing the R2 variation of the LCC inversion model. Finally, the spectral range in the LCC estimation that is less affected by the bandwidth and achieves the optimal inversion accuracy is determined. The results show that the bandwidth and position can significantly affect the inversion accuracy of LCC. The optimal inversion accuracy can be obtained at 575 nm when using single‐band inversion LCC. When using the two‐band vegetation indices, the optimal band positions that are not sensitive to the change of bandwidth are 850–900 nm and 720–740 nm. This study provides a reference for selecting the appropriate bandwidth and position in the LCC inversion.

Abstract The use of computer vision and deep learning in boundary documentation for land registration stems from the ongoing demand for appropriate mapping approaches of unregistered land rights to eradicate the global challenge of tenure insecurity. Previous research has yielded promising results towards automated extraction of photo‐visible cadastral boundaries from high‐resolution imagery. Nonetheless, the extraction of invisible cadastral boundaries is still a challenge. This study investigates the place of sensor/s on‐board unmanned aerial vehicles and deep learning algorithms in detecting cadastral boundaries. It develops a participatory boundary marking procedure using low‐cost markers to bring monument to previously invisible and ill‐defined cadastral boundaries. After that, the researchers trained and tested the accuracy of a convolutional neural network, namely single shot multi‐box detector (SSD) based on Residual Neural Network (ResNet) and Visual Geometry Group (VGG) backbone networks to automatically detect cadastral boundary markers from unmanned aerial vehicle imagery. SSD based on ResNet34 performed best with 0.88 precision, 0.92 recall and 0.91 F measure or (F1) score. VGG19‐based SSD yielded a precision of 0.47, recall of 0.53 and F1 score of 0.50. The horizontal accuracy of the cadastral map generated varied from 0.089 to 0.496 m per parcel, with a standard deviation of 0.120 m. Results show that this approach is practical for cadastral mapping in rural areas.

at the age of 94.With his passing, photogrammetry and remote sensing has lost a brilliant scientist and science manager who was highly respected around the globe and tirelessly active, far beyond his retirement.Gottfried Konecny was a visionary who, in the second half of the 20th century, significantly shaped his field worldwide and who is considered the European father of remote sensing.From the very beginning, his professional life was centred on the sustainable development of our environment.He was also a practitioner, a manager and a diplomat with the ability and determination to achieve his goals.The foundation of his great success was no doubt his ability to not only set his visionary goals but to also achieve them.Gottfried Konecny was born on 17 June 1930 to German parents in Troppau, originally a small Moravian town with a diverse history and a population of around 30,000 at that time, now part of the Czech Republic.His youth in this multilingual region was characterised by the events before and during the Second World War, followed by flight