一种基于局部K-分布的新的SAR图像舰船检测算法

A new ship detection algorithm based on local K-distribution in SAR images

提出了一种基于局部K-分布的新的SAR图像舰船检测算法.取目标窗口和背景窗口,通过把泄露到背景窗口中的舰船部分去除,对背景窗口中的剩余部分统计均值和方差,最终得到杂波分布概率模型进行恒虚警检测.相对于K-分布CFAR检测算法和基于局部窗口的K-分布CFAR检测算法,该算法能够适应杂波的局部变化,对距离很近的舰船不会产生漏检.仿真结果表明了方法的有效性.

The new ship detection algorithm based on local K-distribution proposed in this paper uses a target window and a background window,removes the leaked ship pixels in the background window by using special methods,and estimates the remaining pixels in the background window to get the local clutter’s gray mean and variance. Then the local clutter’s gray probability distribution can be modeled to realize CFAR detection. Compared with K-distribution CFAR detector and local K-distribution CFAR detector,the new algorithm can fit the local change of the sea clutter and detect the targets which are too near. The simulation results show the effectiveness of the new algorithm.