Study On Ships Target Rapid Detection Algorithms With Spaceborne SAR Imagery

It is a serious threat to the legal right and interests of ocean of our country that the ships of several neighboring countries enters the waters under the administration of our country to work on the mapping, monitoring and other illegal activities at present. Therefore, developing the marine target surveillance monitoring technology and improving the ability of marine law enforcement are very essential to the marine security of china. Having the merit of the all-time, all-weather, multi-polar, multi-band and high-resolution imaging capability, Synthetic Aperture Radar has become one of the effective ocean remote sensing monitoring methods. The processed SAR data are always very huge in business applications, so the efficiency of the detection algorithm directly affects the response capability of marine law enforcement. However, domestic and foreign scholars’research work is still rare in the detection efficiency of the algorithm at present. Therefore, this thesis is aiming to explore the methods to fast detect SAR ships. And the main work in the thesis is as follows:1) The data is collected from real experiment. As the main experiment researcher, I have taken part in three SAR and AIS ship synchronization detection experiments, and obtained the real SAR and AIS data. Those data can provide real test data and validate data for subsequent research.2) The comparative analyses are provided for typical SAR ship detection algorithms. On the basis of obtaining the real SAR and AIS data, some classical ship detection algorithms are reproduced and analyzing in comparison with K-CFAR, G0-CFAR, dual parameters CFAR, wavelet transform and KSW. Compared the efficiency and results of each detection algorithm, the optimum algorithm for different SAR images was come out, and provided a strong reference for the practical application of algorithm selection. Meanwhile, the thesis explored the key factors that affect the efficiency of detection, which has laid a solid foundation for developing a superior performance and rapid detection algorithm in the future.3) On the basis of analyzing and grasping the classical SAR ship detection algorithms, a fast SAR ship detection algorithm based on H-CFAR is proposed, which has more practical value. This algorithm directly uses ship SAR image histogram information, and combines the theory of CFAR to make fast and efficient ship target detection come true. The experimental results show that this method effectively improves the efficiency of detection on the premise of assuring detection accuracy, compared with the classical K-CFAR algorithm, and provides an important support for the ocean SAR target real-time surveillance monitoring.