Research on detection and recognition method of large area fixed target at sea based on multisource satellite data

Remote sensing monitoring technology is widely used in the field of maritime target monitoring and supervision. However, the common remote sensing monitoring methods have the problems of large amount of calculation and low efficiency, and there is still some room for improvement in the engineering application of large areas. In view of the above problems, this paper proposes a fast detection and recognition of large-area fixed targets (windmills, offshore platforms, etc.) based on multi-source remote sensing satellite data such as radar and visible light, combined with deep learning technology. Firstly, the multi-phase radar satellite data is used to segment the sea target, and the fixed target is obtained by spatial superposition analysis. Secondly, the fixed target data set of remote sensing image is constructed, and YOLOv5 is used as the network model to carry out the recognition of fixed targets at sea. The experimental results show that in the extraction of fixed targets at sea, the accuracy rate is 100 %, and the accuracy rate of target recognition is 93.18 %.