Xiliang Chen

Few-shot remote sensing scene classification aims to accurately recognize unseen scene categories using only a scarce number of labeled samples, which has emerged as a research hotspot in the field of remote sensing image interpretation. However, remote sensing images intrinsically suffer from large intra-class variations, high inter-class similarities, and complex background interferences. Traditional few-shot learning methods typically perform feature metric learning in Euclidean space, making it difficult to capture the non-Euclidean geometric distribution characteristics of remote sensing features, and they often neglect the spatial structural information embedded in feature maps. To address these issues, this paper proposes a novel few-shot remote sensing scene classification method, termed ZSFS-RGBN, which integrates a Zigzag Scanning Feature Sequence with a Riemannian Geometric Barycenter Network. Specifically, ResNet12 is first employed as the backbone to extract deep convolutional feature maps from both the support and query sets. Second, a Zigzag scanning strategy is introduced to reorganize the two-dimensional feature maps into one-dimensional feature sequences, thereby effectively preserving the spatial locality and structural continuity of the features. Third, an autoregressive moving average (ARMA) model is constructed to characterize the spatial dependencies of the feature sequences, and its state parameters are mapped onto a symmetric positive definite (SPD) matrix manifold, enabling the deep semantic representations of remote sensing scenes in a non-Euclidean geometric space. Finally, a Riemannian geometric barycenter network is designed to learn the Riemannian barycenter of each category on the SPD manifold, where a joint loss function is introduced to simultaneously optimize intra-class compactness and inter-class separability. Comprehensive experiments are conducted on three public remote sensing scene datasets: NWPU-RESISC45, UC Merced Land-Use, and WHU-RS19. Experimental results demonstrate that the proposed method consistently outperforms several representative state-of-the-art approaches under both 5-way 1-shot and 5-way 5-shot settings. Furthermore, ablation studies verify the effectiveness of each component within the proposed framework.

Few-shot remote sensing image scene classification aims to use annotated data to achieve efficient classification, which has important applications in environmental monitoring, urban planning, agricultural yield estimation and other fields. However, due to the significant intra-class variability and inter class similarity in remote sensing scenes, traditional methods struggle to simultaneously capture the spatial correlation and discriminability of features, resulting in limited performance in few-shot scenarios. This paper proposes a few-shot remote sensing image scene classification method based on zigzag scanning feature sequence and Riemannian kernel Fisher discriminant analysis (ZSFS-RKFDA). Firstly, deep features are extracted through a convolutional neural network (CNN), and a zigzag scanning (ZS) process reorganizes the two-dimensional feature maps into one-dimensional sequences that preserve spatial relationships. Secondly, the feature sequence is mapped to non-Euclidean geometric space through the Riemannian manifold space embedding module (RieMSEM). The main steps include constructing an autoregressive moving average (ARMA) model to capture the spatial correlation of features and constructing class prototypes on the Riemannian manifold space. Finally, Riemannian kernel Fisher discriminant analysis (RieKFDA) is employed to optimize the metric distance, enhancing intra-class compactness and inter-class separability. A joint loss function is designed to enable end-to-end training. Experiments on three public datasets demonstrate that the proposed method outper forms existing mainstream approaches in both 5-shot and 10-shot few-shot classification scenarios.

With the development of remote sensing technologies, the volume of remote sensing data is growing dramatically, making efficient management and retrieval of large-scale remote sensing images increasingly important. Recently, deep hashing for content-based remote sensing image retrieval (CBRSIR) has attracted significant attention due to its computational efficiency and high retrieval accuracy. Although great advancements have been achieved, the imbalance between easy and difficult image pairs during training often limits the model’s ability to capture complex similarities and degrades retrieval performance. Additionally, distinguishing images with the same Hamming distance but different categories remains a challenge during the retrieval phase. In this paper, we propose a novel deep relevance hashing (DRH) for remote sensing image retrieval, which consists of a global hash learning model (GHLM) and a local hash re-ranking model (LHRM). The goal of GHLM is to extract global features from RS images and generate compact hash codes for initial ranking. To achieve this, GHLM employs a deep convolutional neural network to extract discriminative representations. A weighted pairwise similarity loss is introduced to emphasize difficult image pairs and reduce the impact of easy ones during training. The LHRM predicts relevance scores for images that share the same Hamming distance with the query to reduce confusion in the retrieval stage. Specifically, we represent the retrieval list as a relevance matrix and employ a lightweight CNN model to learn the relevance scores of image pairs and refine the list. Experimental results on three benchmark datasets demonstrate that the proposed DRH method outperforms other deep hashing approaches, confirming its effectiveness in CBRSIR.

Given the increasingly severe destruction of wetlands in recent years, research and monitoring for wetland protection are urgently needed. However, wetland monitoring still faces significant challenges such as high manual costs and low image monitoring accuracy, primarily due to the complex composition of wetland terrain and diverse spectral characteristics, especially located in remote suburban areas. With their rich semantic and spectral features, studying wetlands as scenes in remote sensing images and segmenting them effectively can address these challenges. To address the issue of low segmentation accuracy on remote sensing image with GrabCut, this study proposes a remote sensing image segmentation extraction method, named SceneCut, based on feature extraction and multifeature joint graph cuts. The method consists of three parts: feature extraction, feature fusion, and multifeature joint graph cuts. In the first part, image features are calculated using a sliding window approach. In the second part, region covariance calculation and matrix eigenvalue decomposition are used to generate a multichannel feature image from the extracted feature vectors. Finally, in the third part, GrabCut is applied to the multichannel image to perform graph cuts, considering multiple features, and the segmentation result is mapped back to the original image to generate the segmented image. Experimental analysis and field exploration validation of wetland images conducted in this study demonstrate that compared to algorithms that only use RGB features for segmentation or those that do not consider the relationship between multiple features, SceneCut performs well in wetland scene extraction, especially in extracting boundaries with significant blending.

In urban scenes, there are man-made ground objects with complex structures and significant height differences, which leads to challenges in generating large-scale true digital orthophoto maps (TDOMs) of cities. However, traditional TDOM generation methods rely on precise prior geometric information, and this reliance makes it difficult to adapt to such complex environments. In this work, we propose TDOM-NeRF, a novel neural radiance field (NeRF) based method for large-scale urban TDOM generation. Without relying on additional prior information, our method utilizes multiview unmanned aerial vehicle (UAV) images as input and implicitly expresses the scene using hash grid features and multilayer perceptrons. By performing orthogonal volume rendering on the reconstruction results of the scene, we effectively address the problem of uneven scales in the synthetic views when generating TDOM. In actual training, we adopt a scene-block training strategy to extend our method to TDOM generation of large-scale scenes while achieving high-fidelity scene reconstruction. In addition, we regularize the model during training to eliminate erroneous “floating objects” in scene reconstruction, making our method applicable to low-overlap UAV datasets. Experimental results demonstrate that our method generates TDOMs with comparable geometric accuracy and eliminates issues such as ghosting, misalignment, and distortion in challenging areas, compared to commercial photogrammetry software like Metashape and Pix4DMapper. Furthermore, our method exhibits superior novel view synthesis performance compared to advanced NeRF methods like Ortho-NeRF. Our pipeline offers a flexible configuration of scene blocking, enabling a balance between performance and efficiency, which makes it particularly suitable for large-scale scene reconstruction.

Few-shot learning (FSL) is a method that does not require a large number of labeled samples and has ability to identify unseen samples. Recently, many excellent FSL methods have been developed for remote sensing scene classification (RSSC). However, for complex remote sensing scenes, the following problems have not been well addressed: 1) the order of local features and the spatial correlation between features haven’t been taken into consideration, 2) the current metric methods are mainly based on Euclidean space (ES), and discriminative ability is limited. Hence, we propose a framework for RSSC with FSL by combining Hilbert feature sequence and Lie group metric space (HFS-LGMS). Specifically, we adopt a lightweight 4-layers convolutional neural network (CNN) as a feature extractor. For the last feature map, we regard each pixel as a multi-dimensional local feature. Then we organize all the features into a feature sequence along the Hilbert curve. To represent the correlation among features, we introduce the autoregressive moving average model (AMAM) to model the observed features and their spatial states. Since the parameters of the model are matrices, which are directly processed as vectors will affect performance. Therefore, we transform the matrix parameter space into the Lie group manifold space (LGMS), and introduce the Lie group intrinsic mean (LGIM) to represent the prototype. The distance between the Lie groups is measured by the geodesic distance. The similarity is measured on Lie group space. Finally, the effectiveness of HFS-LGMS is demonstrated on three public datasets.

Deep neural networks driven by large amounts of annotated samples have been successfully applied in remote sensing scene classification (RSSC). On most common public datasets, the accuracies of scene recognition tasks have been close to saturation. A learning paradigm that specifically tackles this issue has emerged, i.e., few-shot learning (FSL). FSL methods for natural image recognition are developing rapidly. Recently, these methods have been widely used in RSSC. Compared with natural images, the problems of intra-class differences and inter-class similarities in RSSC are more serious, which hinders the further development of FSL methods for RSSC. To address this issue, we propose a multi-manifold metric learning framework for RSSC with FSL. Specifically, we use a lightweight convolutional neural network as a feature extraction block. To further enhance the representation ability of features, we embed the feature map into two heterogeneous and complementary Riemannian manifold geometric structures, i.e. Grassmannian manifold and symmetric positive definite (SPD) manifold. Then, to facilitate the measurement, we introduce the Riemann kernel function to embed two heterogeneous manifold structures into the high-dimensional Hilbert space for fusion. Finally, we design a learnable distance metric scheme that can be optimized according to the divergence of inter-class pairs and intra-class pairs, thereby reducing the impact of large intra-class differences and high inter-class similarity on scene recognition. We verify the effectiveness of MMML on three commonly used datasets. The results show that the accuracy of our proposed method is from 1.21% to 3.12% higher than the state-of-the-art methods.

In recent years, supervised learning, represented by deep learning, has shown good performance in remote sensing image scene classification with its powerful feature learning ability. However, this method requires large-scale and high-quality handcrafted labeled datasets, which leads to a high cost of obtaining annotated samples. Self-supervised learning can alleviate this problem by using unlabeled data to learn the image’s feature representation and then migrate to the downstream task. In this study, we use an encoder–decoder structure to construct a self-supervised learning architecture. In the encoding stage, the image mask is used to discard some of the image patches randomly, and the image’s feature representation can be learned from the remaining image patches. In the decoding stage, the lightweight decoder is used to recover the pixels of the original image patches according to the features learned in the encoding stage. We constructed a large-scale unlabeled training set using several public scene classification datasets and Gaofen-2 satellite data to train the self-supervised learning model. In the downstream task, we use the encoder structure with the masked image patches that have been removed as the backbone network of the scene classification task. Then, we fine-tune the pre-trained weights of self-supervised learning in the encoding stage on two open datasets with complex scene categories. The datasets include NWPU-RESISC45 and AID. Compared with other mainstream supervised learning methods and self-supervised learning methods, our proposed method has better performance than the most state-of-the-art methods in the task of remote sensing image scene classification.