高分辨SAR图像中杂波的统计特性分析

Statistical Analysis of Clutter in High-Resolution SAR Images

为使SAR图像目标检测在复杂多变的地物杂波环境中获得最佳的性能,研究SAR图像中杂波的统计特性将是一件有意义的工作。在综合对比已有统计模型的基础上,从理论上重点分析了G^0分布。运用实测SAR场景中表征不同地物类型的大量数据,从幅度直方图拟合、拟合精度检脸等方面深入分析了杂波的统计特性。得出了在目前已有实用的统计模型中,G^0分布最适合于描述SAR图像杂波统计特性,它是一种对于均匀、一般不均匀和极不均匀杂波区域建模的通用分布的新结论。因此,以C^0分布为杂波模型的目标检测算法具有巨大的应用潜力。

In order to ensure the best performance of target detection in SAR images in complex and varying clutter,it is significant to study the statistical property f clutter. Based on the synthetic comparison of the existing statistical model ,the Go distribution is analyzed in theory with emphasis. By applying the real SAR data denoting different terrain categories, the clutter is analyzed comprehensively with amplitude histogram fitting,the goodness-of-fit, etc. A new conclusion is obtained that the Go distribution is most appropriate to describe the statistical property of clutter and is a general distribution for statistically modeling the homogeneous,heterogeneous and extremely heterogeneous clutter regions. Thus ,the Go distribution has great potential to the algorithm design of target detection.