SAR图像舰船目标双阈值快速CFAR检测算法

A double threshold fast CFAR ship detection algorithm in SAR images

针对舰船检测的高效性现实需求,本文提出了改进的舰船目标CFAR检测算法。在全局阈值的基础上,利用第一虚警率检测出超过全局阈值的像素点,将超过全局阈值的像素点视作疑似目标点,将低于全局阈值的像素点视作杂波点。在局部窗口检测过程中,引入筛选机制,设置第二虚警率,将背景窗口中的疑似像素目标剔除得到最终的检测结果。该算法计算量少,并且改善了多目标环境下的检测性能,消除了泄露到背景窗口中的目标像素对杂波统计特性的影响。仿真结果表明算法的有效性。

Aimed at the demand of efficiency,an improved CFAR detection algorithm for vessels was proposed. After pre-detection based on global threshold,the pixels,which were lower than the global threshold given by the first false alarm rate,would be processed as clutter,the others,which were higher than the global threshold, would be processed as suspected targets. During the local window CFAR detection progress,which was a more accurate step,the second false alarm ratio,which was given by the second false alarm rate,was set to filter the suspected target pixels and get the final result. This method is less amount of calculation and improves the performance of multiple object environments. Besides,this method can also remove the influence for statistics properties of clutters due to the target pixels which leak to the background window. The experiment showed the effectiveness of this method.