Abstract: The classical Parzen algorithmis is based on the assumption that the targets occupy a small part of the SAR image and uses all pixels of the SAR image to estimate the probability density function of the clutter background.This method results in the elevation of the detection threshold,and then it is possible to miss targets that are less obvious.In order to overcome this problem,we proposed an improved Parzen detection algorithm.The proposed algorithm adaptively sets the target windows according to the size of the target and deletes the potential targets from the background.Then it estimates the clutter distribution based on Parzen window method.Finally,it determines the detection threshold for the target detection.Compared with the traditional Parzen detection method,the proposed algorithm decreases the number of missing target and also improves the detection performance.The detection results with the real SAR images verify the effectiveness of the algorithm.

Key words: SAR image,Ship,Detection,Parzen windows