首页 > , Vol. , Issue () : -
With the rapid development of spatial technology, the resolution of remote sensing images is gradually im-proved, and the detailed information and spatial information contained in remote sensing images are also richer, which leads to the problem that the difference between different categories becomes smaller and the difference between the same categories becomes larger, that is, the phenomenon of the same spectrum of foreign objects and the different spectrum of the same objects is serious. However, the existing dual-modal segmentation methods do not extract the dual-modal feature information of remote sensing images sepa-rately, and the fusion features are not sufficient, and the details of upsampling recovery are insufficient, resulting in the inability to accurately and efficiently learn remote sensing image information, resulting in segmentation errors, edge blur and other problems. This study proposes a high-resolution remote sensing image segmentation based on dual-modal effi-cient feature learning. The algorithm designs appropriate encoders for different modal remote sensing im-ages, efficiently extracts dual-modal features, and reduces the differences between different path features through interactive reinforcement modules. Then, the dual-modal feature aggregation module and the deep feature extraction module are proposed to further fuse and extract the dual-modal features, so that the net-work can fully learn the complementary information of the dual-modal. Finally, a multi-layer feature up-sampling module is proposed, which uses the high-level features with rich semantic information to weight the low-level features with rich detail information, and gradually upsampling to achieve efficient feature recovery and improve segmentation performance. In this paper, experimental on the Potsdam and Vaihingen datasets demonstrate the overall accuracy reaches 94.52 % and 90.45 % respectively. The experimental results show that the segmentation effect of the proposed algorithm is better than that of the existing algorithms, which can efficiently extract and fuse the multi-modal complementary features of high-resolution remote sensing images, and improve the seg-mentation accuracy of remote sensing images. This study proposes a high-resolution remote sensing image segmentation based on dual-modal effi-cient feature learning. Experiments on ISPRS Potsdam and Vaihingen datasets show that the proposed model is more suitable for segmenting low vegetation and trees, buildings and roads with very similar spectral features, and can also achieve accurate segmentation of small targets such as cars. However, the complexity of the model needs to be further reduced, and there is still much room for improvement in ac-curacy. In the future, a better segmentation network will be designed to fuse more than two modal features to obtain more feature information to achieve more accurate remote-sensing image segmentation.