Deep convolution neural network (DCNN) technology has achieved great success in extracting buildings from aerial images. However, the current mainstream algorithms are not satisfactory in feature extraction and classification of homesteads, especially in complex rural scenarios. This study proposes a deep convolutional neural network for rural homestead extraction consisting of a detail branch, a semantic branch, and a boundary branch, namely Multi-Branch Network (MBNet). Meanwhile, a multi-task joint loss function is designed to constrain the consistency of bounds and masks with their respective labels. Specifically, MBNet guarantees the details of prediction through serial 4× down-sampled high-resolution feature maps and adds a mixed-scale spatial attention module at the tail of the semantic branch to obtain multi-scale affinity features. At the same time, the low-resolution semantic feature maps and interaction between high-resolution detail feature maps are maintained. Finally, the result of semantic segmentation is refined by the point-to-point module (PTPM) through the generated boundary. Experiments on UAV high-resolution imagery in rural areas show that our method achieves better performance than other state-of-the-art models, which helps to refine the extraction of rural homesteads. This study demonstrates that MBNet is a potential candidate for building an automatic rural homestead management system.

It is very significant for rural planning to accurately count the number and area of rural homesteads by means of automation. The development of deep learning makes it possible to achieve this goal. At present, many effective works have been conducted to extract building objects from VHR images using semantic segmentation technology, but they do not extract instance objects and do not work for densely distributed and overlapping rural homesteads. Most of the existing mainstream instance segmentation frameworks are based on the top-down structure. The model is complex and requires a large number of manually set thresholds. In order to solve the above difficult problems, we designed a simple query-based instance segmentation framework, QueryFormer, which includes an encoder and a decoder. A multi-scale deformable attention mechanism is incorporated into the encoder, resulting in significant computational savings, while also achieving effective results. In the decoder, we designed multiple groups, and used a Many-to-One label assignment method to make the image feature region be queried faster. Experiments show that our method achieves better performance (52.8AP) than the other most advanced models (+0.8AP) in the task of extracting rural homesteads in dense regions. This study shows that query-based instance segmentation framework has strong application potential in remote sensing images.

The basic information survey on homesteads requires understanding the shape of homesteads, and the shape of the homesteads based on the spatial location can reflect information such as their outline and regularity, but the current shape classification of rural homesteads at the parcel scale lacks analytical methods. In this study, we endeavor to explore a classification model suitable for characterizing homestead shapes at the parcel scale by assessing the impact of various research methods. Additionally, we aim to uncover the evolutionary patterns in homestead shapes. The study focuses on Yangdun Village, located in Deqing County, Zhejiang Province, as the research area. The data utilized comprise Google Earth satellite imagery and a vector layer representing homesteads at the parcel scale. To classify the shapes of homesteads and compare classification accuracy, we employ a combination of methods, including the fast Fourier transform (FFT), Hu invariant moments (HIM), the Boyce and Clark shape index (BCSI), and the AlexNet model. Our findings reveal the following: (1) The random forest method, when coupled with FFT, demonstrates the highest effectiveness in identifying the shape categories of homesteads, achieving an average accuracy rate of 88.6%. (2) Combining multiple methods does not enhance recognition accuracy; for instance, the accuracy of the FFT + HIM combination was 88.4%. (3) The Boyce and Clark shape index (BCSI) proves unsuitable for classifying homestead shapes, yielding an average accuracy rate of only 58%. Furthermore, there is no precise numerical correlation between the homestead category and the shape index. (4) It is noteworthy that over half of the homesteads in Yangdun Village exhibit rectangular-like shapes. Following the “homesteads reform”, square-like homesteads have experienced significant vacating, resulting in a mixed arrangement of homesteads overall. The research findings can serve as a methodological reference for the investigation of rural homestead shapes. Proficiency in homestead shape classification holds significant importance in the realms of information investigation, regular management, and layout optimization of rural land.

Precise and timely classification of land cover types plays an important role in land resources planning and management. In this paper, nine kinds of land cover types in the acquired hyperspectral scene are classified based on the kernel collaborative representation method. To reduce the spectral shift caused by adjacency effect when mining the spatial-spectral features, a correlation coefficient-weighted spatial filtering operation is proposed in this paper. Additionally, by introducing this operation into the kernel collaborative representation method with Tikhonov regularization (KCRT) and discriminative KCRT (DKCRT) method, respectively, the weighted spatial-spectral KCRT (WSSKCRT) and weighted spatial-spectral DKCRT (WSSDKCRT) methods are constructed for land cover classification. Furthermore, aiming at the problem of difficulty of pixel labeling in hyperspectral images, this paper attempts to establish an effective land cover classification model in the case of small-size labeled samples. The proposed WSSKCRT and WSSDKCRT methods are compared with four methods, i.e., KCRT, DKCRT, KCRT with composite kernel (KCRT-CK), and joint DKCRT (JDKCRT). The experimental results show that the proposed WSSKCRT method achieves the best classification performance, and WSSKCRT and WSSDKCRT outperform KCRT-CK and JDKCRT, respectively, obtaining the OA over 94% with only 540 labeled training samples, which indicates that the proposed weighted spatial filtering operation can effectively alleviate the spectral shift caused by adjacency effect, and it can effectively classify land cover types under the situation of small-size labeled samples.