Research on Segmentation Method of Remote Sensing Image Based on Multimodal Fusion

With the development of remote sensing technology, remote sensing image segmentation methods have been widely used in the fields of geographic information acquisition, environmental monitoring, and disaster warning. However, the traditional single-modal image segmentation methods often face the problem of insufficient accuracy when dealing with complex background and diverse scenes. In this paper, a remote sensing image segmentation method based on multi-modal fusion is studied, which aims to solve the above problems. By combining the information of optical image and infrared image, the attention mechanism and feature-level fusion strategy in deep learning network are used to enhance the segmentation performance of the model in complex environment. Experiments on WHU building dataset and Vaihingen dataset show that the proposed method outperforms other classical networks in several performance metrics, showing higher segmentation accuracy and robustness.