基于Otsu和模糊核聚类算法的极化SAR图像分类

Classification of Polarimetric Synthetic Aperture Radar Images Based on Otsu's Method and Fuzzy Kernel C-means Clustering Algorithm

随着极化技术的发展,越来越多的高分辨率SAR图像出现,对其分类成为一个耗时的工作。针对该问题,提出了基于Otsu和模糊核聚类算法的极化SAR图像分类。将SAR图像分类分为两步,第一步借助Otsu法进行粗略分割,将图像转化为若干个区域;第二步利用模糊核聚类算法将第一步剩下的像素进一步分类。实验结果表明,该算法既可保持较高的分类精确度,又可保证较快的计算速度。

With the development of polarimetric technology, more and more high resolution SAR images are ob- tained, and classification of these images is time consuming. A modified algorithm combined Otsu＇s method and fuzzy kernel c-means algorithm is proposed. The algorithm is implemented by two steps. Firstly, the SAR image is segmented into some blocks coarsely by Otsu＇s method. Then the processed SAR image is further classified by fuzzy kernel c-means algorithm. The algorithm can not only has better precision, but also get higher calculating speed.