一种基于多模分布的SAR图像分割算法

摘要

In SAR image, the different scattering mechanisms of target, background clutter and target shadow are different , thus resulting in their different statistical characteristics. By analyzing their respective imaging characteristics, different statistical models are developed to describe the three components. Then a multi-mode distribution SAR image segmentation algorithm based on a combination of the three models is proposed. For the target and shadow segmentation, different pre-processing methods are adopted. A fast one-dimensional Otsu algorithm is proposed to segment target from background clutter; and a SAR image enhancement algorithm based on the background mean-preserving shock filter is proposed to segment shadow from background clutter. Experimental results show that, compared with traditional segmentation algorithms based on the single-mode distributions, this multi-mode distribution segmentation algorithm can make better use of the statistical characteristic differences of the three components, thus obtaining higher segmentation accuracy and better practicability.%SAR图像目标、背景、阴影的不同散射机理使得这三部分具有不同的统计特性.分析SAR图像三部分统计性质分别建立统计模型,并给出了一种基于这三种模型组合的多模分布的SAR图像分割算法,对于目标分割和阴影分割分别采用不同的预处理方法,分别提出了快速Otsu分割算法分割目标和基于背景均值保留的冲击滤波算法分割阴影.分割结果表明,这种基于多模分布的SAR图像分割算法与传统的基于单模分割算法相比,能最佳地利用各部分统计特性的差异,得到准确的分割结果.