基于邻域统计分布变化分析的UWB SAR隐蔽目标变化检测

摘要

Because of large pixel value change between multitemporal UWB SAR images caused by different imaging geometries, the performance of change detection algorithm based on pixel value difference declines quickly.In order to deal with this problem, a new UWB SAR foliage target change detection algorithm based on local statistic distribution is proposed. In the algorithm, the Gram-Charlier expansion theory and rank order filter are combined to estimate local statistic distribution. Then, the K-L divergence is used to measure the change between local statistic distribution of multitemporal UWB SAR image. And the target can be detected because of large K-L divergence value. Finally, the experimental results show that the algorithm can better deal with the pixel value change between multitemporal UWB SAR images with different imaging geometries and an obvious performance improvement on detection can be obtained.%该文针对载机不同航迹条件下所得多时相UWB SAR图像灰度值存在较大起伏,严重影响了基于像素灰度值差异的变化检测算法性能,提出了一种基于邻域统计分布变化分析的UWB SAR隐蔽目标变化检测方法.该方法将Gram-Charlier展开理论同秩序滤波器相结合对多时相图像中每个像素邻域的统计分布进行估计,进而借助K-L散度理论对多时相图像邻域统计分布变化进行定量分析以检测目标对应的变化区域.实验结果表明,该文方法能够更好地适应不同航迹UWB SAR图像间灰度起伏的影响,取得更好的检测结果.