Algorithm Research Of Ship Target Detection For Synthetic Aperture Radar

Since a vast sea area of our country,a large number of ships and an increasing demand for safe navigation,it is of great significance to improve the detection accuracy of the ship targets in order to safeguard our country’s maritime safety.Synthetic Aperture Radar(SAR)has been more and more widely used due to its ability to penetrate into rain,cloud and smoke with the capability of working all day/night in and be capable of imaging sea surface from different angles.This paper explores the detection method of ship target in SAR image based on local contrast measure(LCM)and extended fractal features(EF).Firstly,this paper introduces several common methods of image grizzled processing and filtering method used in image denoising,compares the simulation results,and analyzes the advantages and disadvantages of each method and their own applicable range.Finally select the appropriate image gray processing method and filtering method to be applied to the ship target detection in this paper.The effect of image preprocessing will directly affect the effectiveness and reliability of the subsequent target detection.Secondly,based on the analysis of the advantages and disadvantages of LCM,a large number of false alarm phenomena will be produced when ship targets are around the strong background noise condition.The local contrast measure is redefined,and then an improved local contrast is proposed.The reason of the change is analyzed theory.Through comparison experiments,the results show that the false alarm can be suppressed effectively while guaranteeing the detection efficiency of the algorithm.Finally,the realization process of the traditional extended fractal detection algorithm is introduced,the sensitivity of the target size is analyzed,and then the second order statistic characteristic of EF is discussed.In order to solve the problem of long time and false alarm in the case of a large scale SAR image,this paper proposes an extended fractal in eight direction based on mean down-sampling.It is proved that,in the process of target detection,the down-sampling operation can effectively improve the efficiency of the algorithm,and the improved algorithm can suppress the false alarm in the case of correctly detecting the target and no leakage of the alarm,compared with the traditional two-directional extended fractal.