分析了合成孔径雷达(SAR)图像的统计分布特性,推导了SAR目标及阴影的分割阈值与虚警概率之间的关系,给出了一种基于Weibull分布的SAR图像目标及目标阴影的分割方法,首先根据推导出的分割阈值公式对SAR目标图像及阴影图像进行粗分割,然后利用形态学滤波和计数滤波器进行细分割。基于MSTARSAR实测数据进行的大量实验表明,该方法是有效的,优于传统的双参数恒虚警分割方法。
SAR target and shadow segmentation based on Weibull distribution are proposed. First, the statistic property and the distribution parameters estimation of SAR image are analyzed. Then, the relationship between segmental thresholds and CFAR (constant false alarm rate) is derived. The target and its shadow in SAR image are segmented by the thresholds, morphological filter and counting filter respectively. Experimental results with MSTAR (Moving and Stationary Target Acquisition and Recognition ) SAR data sets show the better performance of the proposed method compared with that of the two parameter CFAR.