A Real-time Ship Detection Algorithm for High Resolution SAR Image

SyntheticApertureRadar (SAR) image ship target surveillance is an important aspect of the application of SAR in Marine remote sensing. It has great significance in fishery control, maritime traffic management, combating piracy, protecting the seas and protecting rights and interests. Spaceborne real-time processing technology is one of the important development directions of high-resolution spaceborne SAR in the future. Currently, ship detection of SAR images generally has many problems, such as large computation, inaccurate feature extraction and high false alarm rate, which make it difficult to realize efficient and accurate real-time detection.In this paper, a CFAR (Global Constant False-Alarm Rate) SAR image ship target detection algorithm is proposed to solve the problem of inaccurate feature extraction caused by large computation amount of ship detection and serious sidelobe in high resolution SAR image. Firstly, based on the Gaofen-3 image, the global CFAR algorithm is used for preliminary detection. Then, the connected region is extracted, and ROI is selected by the number of pixels in the connected region. Then, the influence of ROI sidelobe was removed by an iterative rotation sidelobe removal method, and its geometric features were extracted. Finally, the geometric features are used to screen out the final detection results. The simulation results show that the proposed algorithm has better efficiency and detection performance than the traditional CFAR method.