Accurate extraction of buildings using high spatial resolution imagery is essential to a wide range of urban applications. However, it is difficult to extract semantic features from a variety of complex scenes (e.g., suburban, urban and urban village areas) because various complex man-made objects usually appear heterogeneous with large intra-class and low inter-class variations. The automatic extraction of buildings is thus extremely challenging. The fully convolutional neural networks (FCNs) developed in recent years have performed well in the extraction of urban man-made objects due to their ability to learn state-of-the-art features and to label pixels end-to-end. One of the most successful FCNs used in building extraction is U-net. However, the commonly used skip connection and feature fusion refinement modules in U-net often ignore the problem of feature selection, and the ability to extract smaller buildings and refine building boundaries needs to be improved. In this paper, we propose a trainable chain fully convolutional neural network (CFCN), which fuses high spatial resolution unmanned aerial vehicle (UAV) images and the digital surface model (DSM) for building extraction. Multilevel features are obtained from the fusion data, and an improved U-net is used for the coarse extraction of the building. To solve the problem of incomplete extraction of building boundaries, a U-net network is introduced by chain, which is used for the introduction of a coarse building boundary constraint, hole filling, and "speckle" removal. Typical areas such as suburban, urban, and urban villages were selected for building extraction experiments. The results show that the CFCN achieved recall of 98.67%, 98.62%, and 99.52% and intersection over union (IoU) of 96.23%, 96.43%, and 95.76% in suburban, urban, and urban village areas, respectively. Considering the IoU in conjunction with the CFCN and U-net resulted in improvements of 6.61%, 5.31%, and 6.45% in suburban, urban, and urban village areas, respectively. The proposed method can extract buildings with higher accuracy and with clearer and more complete boundaries.

Scene classification plays a significant role in the field of remote sensing (RS). Recently, the rapid development of convolutional neural networks (CNNs) has enabled a vital breakthrough in high-resolution RS image scene classification. However, complex backgrounds and small objects in high-resolution RS images pose challenges to the application of CNNs. To this end, a novel multi-branch fusion attention network (MBFANet) is proposed to improve the feature extraction ability and generalization performance of models. Specifically, a multi-branch fusion attention module (MBFAM) is designed by adaptively fusing two parallel submodules, namely efficient pooling channel attention module (EPCAM) and efficient convolution coordinate attention module (ECCAM), which helps the model focus on more key cues in images that are difficult to classify. The ablation experiments in RS scene classification datasets demonstrate the effectiveness of our methods. In addition, MBFANet achieves competitive results on three benchmark datasets.

Timely and accurate information on rural settlements is essential for rural development planning. Remote sensing has become an important means for accurately mapping large scale rural settlements. Nevertheless, numerous difficulties remain in accurate and efficient rural settlement extraction. In this study, by combining multi-dimensional features derived from Sentinel-1/2 images, Visible Infrared Imaging Radiometer Suite supporting a Day-Night Band (VIIRS-DNB) dataset, and Digital Elevation Model (DEM) data using the Google Earth Engine (GEE) platform, we proposed an efficient framework with good transferability for mapping rural settlements in the Yangtze River Delta. To avoid the time-consuming selection of a large number of training samples in the whole study area, we employed four random forest models obtained from the training samples in respective training municipal districts in four different regions to classify other municipal districts in their corresponding region. We found that different features play diverse vital roles in the extraction of rural settlements in various regions. Compared to results only using optical data, accuracies obtained by the proposed method were significantly improved. The average user’s accuracy, producer’s accuracy, overall accuracy, and Kappa coefficient increased by 16.75%, 17.75%, 11.50%, and 14.50% in the four training municipal administrative areas, respectively. The overall accuracy and Kappa coefficient were 96% and 0.84, respectively. By contrast, our classification results are superior to other public datasets. The final mapping results provided a detailed spatial distribution of the rural settlements in the Yangtze River Delta and revealed that the total area of rural settlements is approximately 32,121.1 km², accounting for 17.41% of the total area. The high-density rural settlements are mainly distributed in the Northern Plain and East Coast, while the low-density rural settlements are located in the Central Hills and Southern Mountain.

Remote sensing (RS) scene classification is challenging due to changes in the scale and direction of scenes within a category. Bilinear pooling method can extract higher-order and spatial orderless information and has been shown to achieve impressive performance on various visual tasks. However, bilinear pooled features are high dimensional, which makes them impractical for subsequent processing, especially for the convolutional neural network (CNN) models with more channels in the final convolutional layer. To alleviate this shortcoming, an improved bilinear pooling method is proposed to build the compact bilinear CNN model in this work. Specifically, a joint pooling method is proposed to reduce the high-dimensional bilinear features, and it can be embedded in a bilinear CNN architecture for end-to-end optimization. Through the experimental evaluation of three real RS scene image data sets, it is proved that the improved bilinear pooling method can obtain features with higher discriminative power than the bilinear pooling method but with lower dimensionality. In addition, it also reduces the running time of model training.

As the manual detection of building footprint is inefficient and labor-intensive, this study proposed a method of building footprint extraction and change detection based on deep convolutional neural networks. The study modified the existing U-Net model to develop the “PRU-Net” model. PRU-Net incorporates pyramid scene parsing (PSP) to allow multiscale scene parsing, a residual block (RB) in ResNet for feature extraction, and focal loss to address sample imbalance. Within the proposed method, building footprint extraction is conducted as follows: 1) unmanned aerial vehicle images are cropped, denoised, and semantically marked, and datasets are created (including training/validation and prediction datasets); 2) the training/validation and prediction datasets are input into the full convolutional neural network PRU-Net for model training/validation and prediction. Compared with the U-Net, PSP+U-Net (PU-Net), and U-Net++ models, PRU-Net offers improved footprint extraction of buildings with a range of sizes and shapes. The large-scale experimental results demonstrated the effectiveness of the PSP module for multiscale scene analysis and the RB module for feature extraction. After demonstrating the improvements in building extraction offered by PRU-Net, the building footprint results were further processed to generate a building change map.

Change detection has become a crucial topic in the field of remote sensing deep learning due to its extensive application in earth observation. However, real remote sensing images often contain multiple land cover classes with significant intraclass variability and interclass similarity, limiting the performance of change detection in complex scenarios. To address this, we leverage the capabilities of vision foundation models by applying the segment anything model (SAM) to remote sensing change detection, and we name this method dual encoder–decoder SAM (DED-SAM). Specifically, we construct a DED framework, utilizing a small-scale change detection model in both branches to generate mixed prompts including image features, mask prompts, and box prompts. The SAM 2 model is used for fine-grained recognition of dual-temporal images, generating accurate, and stable feature boundaries, which are then used as constraints to generate the final change mask. To validate the effectiveness of DED-SAM across various application scenarios, we conduct quantitative experiments on three public datasets: Levir-CD, SYSU-CD, and CDD, testing its detection capabilities under single change categories, multiple change categories, and seasonal pseudochange interference. The results show that the proposed DED-SAM achieved state-of-the-art F1 scores and IoUs on these three datasets: LEVIR-CD (92.00%, 85.11%), SYSU-CD (84.15%, 72.01%), and CDD (97.72%, 95.47%).