Research On Ship Detection Algorithm For Spaceborne SAR Image

Marine ship monitoring is of great significance to maritime traffic and national defense security.Spaceborne Synthetic Aperture Radar(SAR)has full-time all-weather imaging capability and has become an important means of sea observation.The SAR image ship detection technology can realize the real-time and accurate positioning of the ship in a complex environment,which has become a hot research topic at home and abroad.In this paper,based on the SAR image ship detection problem,the methods of Constant False Alarm Rate(CFAR)based on Maximally Stable Extremal Regions(MSER)and ship detection based on neural network SAR image are studied.The main contents are as follows:(1)In view of the slow detection speed of traditional CFAR,the advantages of twostage CFAR algorithm and a fast calculation method of parameter estimation compared with traditional CFAR are analyzed from the aspects of algorithm improvement and parameter estimation calculation optimization.(2)Aiming at the slow detection speed of the two-stage CFAR algorithm,a MSERbased CFAR ship detection algorithm is proposed,which can accurately and quickly locate the potential position of the ship,and achieve improved detection speed and detection accuracy.The performance of the MSER-based CFAR is better than the twolevel CFAR algorithm.(3)Aiming at the low adaptability of CFAR-based algorithms,a ship detection algorithm based on Cascaded Deconvolutional Single Shot Detector(DSSD)is proposed.The algorithm can effectively detect complex scenes,and the detection accuracy and adaptability are better than CFAR algorithm.The measured spaceborne SAR data experiments verify the effectiveness of the above algorithm.The results show that the MSER-based CFAR algorithm and the serial DSSD-based algorithm can effectively realize SAR image ship detection,which improves the detection speed and accuracy compared with the traditional CFAR algorithm.