Resolution is a comprehensive reflection and evaluation index for the visual quality of remote sensing images. Super-resolution processing has been widely applied for extracting information from remote sensing images. Recently, deep learning methods have found increasing application in the super-resolution processing of remote sensing images. However, issues such as blurry object edges and existing artifacts persist. To overcome these issues, this study proposes an improved generative adversarial network with self-attention and texture enhancement (TE-SAGAN) for remote sensing super-resolution images. We first designed an improved generator based on the residual dense block with a self-attention mechanism and weight normalization. The generator gains the feature extraction capability and enhances the training model stability to improve edge contour and texture. Subsequently, a joint loss, which is a combination of L1-norm, perceptual, and texture losses, is designed to optimize the training process and remove artifacts. The L1-norm loss is designed to ensure the consistency of low-frequency pixels; perceptual loss is used to entrench medium- and high-frequency details; and texture loss provides the local features for the super-resolution process. The results of experiments using a publicly available dataset (UC Merced Land Use dataset) and our dataset show that the proposed TE-SAGAN yields clear edges and textures in the super-resolution reconstruction of remote sensing images.
The functional classification of buildings is important for creating and managing urban zones and assisting government departments. Building function recognition is incredibly valuable for wide applications ranging from the determination of energy demand. By aiming at the topic of urban function classification, a semi-supervised graph structure network combined unified message passing model was introduced. The data of this model include spatial location distribution of buildings, building characteristics and the information mined from points of interesting (POIs). In order to extract the context information, each building was regarded as a graph node. Building characteristics and corresponding POIs information were embedded to mine the building function by the graph convolutional neural network. When training the model, several node labels in the graph were masked, and then these labels were predicted by the trained model so that this work could take full advantage of the node label and the feature information of all nodes in both the training and prediction stages. Quasi-experiments proved that the proposed method for building function classification using multi-source data enables the model to capture more meaningful information with limited labels, and it achieves better function classification results.
1 April 2024 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Digital roads are a crucial component of smart city development and sustainable urban development, being one of the fundamental geographic elements. Remote sensing images and GPS trajectories are two important data sources for obtaining road information, which can describe roads from complementary perspectives. Either of these data sources has its limitations in road extraction tasks, but their fusion together can enhance road detection performance. However, previous work on the fusion between remote sensing images and GPS trajectories does not fully utilize the information of these two modalities, and both suffer from the problem of information loss. Therefore, we propose a deep convolutional neural network called RCAF-Net that fuses remote sensing images and GPS trajectories to extract road information more efficiently. We designed a residual attention fusion module to fuse the road features of both modalities from multiple layers and scales. In addition, we consider other attribute information, such as the speed and direction of the GPS trajectories, to provide richer semantic information for the road extraction task. The comprehensive experiments on two datasets show that RCAF-Net achieves about a 3% improvement in IoU and F1-score over the earlier data fusion. Our method fuses remote sensing images and GPS trajectories more efficiently than the existing methods and can extract more fine-grained and complete road information.
The classification of urban functional areas is important for understanding the characteristics of urban areas and optimizing the utilization of urban land resources. Existing related methods have improved accuracy. However, they neglect cognitive differences amongst humans in the different scales of regional functions. Moreover, how to build the correlations of cross-scale characteristics is still unresolved when realizing the classification of multiscale urban functional zones. To resolve these problems, a transportation analysis zone involving urban buildings as research units is created and these units are described by geometric and functional characteristics using multiple data sources. Then, a hierarchical clustering model is built for the recognition of urban functional areas at varying scales with landmark semantic constraints. In the experiments, Shanghai served as the study area, and multiscale zones were created using different levels of road networks considering the constraint correlation of the significance between cross-scale maps. The experiential results show the proposed method has excellent performance and optimizes the functional zone classification at different scales. This study not only enriches the multiscale urban functional area-recognition methods but also can be used in other aspects, like cartographic generalization or spatial analysis.
11 September 2025 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Visual Geo-localization (VG) is a critical task of the geolocation field, aiming to match query images with a large database of images from diverse locations. Recent research has shown that pretrained models for general visual learning tasks perform exceptionally well on large datasets, offering effective feature representations for a wide range of visual perception problems. Nevertheless, current models demonstrate constrained generalizability in feature extraction, with their performance in visual grounding tasks deteriorating under complex environmental conditions. This limitation principally arises from three interrelated challenges: sensitivity to illumination fluctuations, vulnerability to perspective alterations, and susceptibility to interference from non-discriminative visual clutter. To address these issues, we propose a new feature aggregation model that combines features into global descriptors focused on discriminative regions by integrating global relationships within each feature map through a cascade of feature blending. In addition, we introduce attention mechanisms that aggregate global information and emphasize task-relevant regions to guide the re-discrimination of local features. The image similarity score is then derived and reordered based on the optimal number of matching points between the candidate images and the query image. Experimental results demonstrate that our method performs well on large-scale datasets across multiple benchmarks. Furthermore, our approach outperforms single-stage methods using general pretrained models and the benchmark of two-stage methods in most cases.
1 January 2026 in IEEE Transactions on Geoscience and Remote Sensing