一种基于SAR图像的典型线特征抽取方法

Typical linear feature extraction based on SAR image

针对合成孔径雷达(Synthetic Aperture Radar)图像的相干斑噪声会使图像的分割和目标的提取困难的问题,文中利用改进的Lee滤波手段去除相干斑噪声,并使用照度均衡算法平衡SAR图像的光照度,最后使用图割(Graph Cuts)算法交互式地对SAR图像中的线性结构进行抽取。实验结果表明Lee滤波和照度均衡算法分别对于抑制相干斑噪声干扰和修正图像对比度有显著效果,图割算法可以有效地提取出用户感兴趣的线性结构。

For SAR(Synthetic Aperture Radar)images extracted difficult questions speckle makes the image segmentation and objectives,this paper uses modified Lee filter means removing speckle and illumination equalization algorithm for the balance of the SAR image illumination,finally using graph cuts(Graph Cuts)algorithm for SAR images interactively linear structure decimated.The results show that Lee filtering achieves balanced and illumination algorithms respectively for suppressing speckle noise and contrast-corrected image has a significant effect,graph cut algorithm can effectively extract the user is interested in a linear structure.