中文摘要：

目的改进模糊C-均值FCM算法，并对SAR图像进行粗、细分类。方法对FCM算法从初始聚类中心、隶属度约束条件两个方面进行改进，并提出对SAR图像的粗、细分类。首先利用改进的FCM算法对图像进行聚类，然后在隶属度矩阵中设定阈值，对小于阈值的像素块进行进一步细分类。结果得到并验证了改进的FCM算法，该算法对图像进行分类的分类精度比传统的FCM算法要高。结论本算法既可以保持较高的精确度，又可保证较快的计算速度。

英文摘要：

Aim In order to improve Fuzzy C-means （FCM） Traditional algorithm and classify SAR images with coarse and fine stages. Methods The FCM algorithm is improved based on the situation of initialization and the constraint on subordinate degree. And the new algorithm is used in SAR image classification by clustering the image with the improved FCM algorithm, and putting a threshold in membership function firstly, and then, doing further fine classification for the pixel window whose subordinate degree is less than the threshold. Results Experimental results show that the precision of classification using the proposed algorithm is better. Conclusion This algorithm can maintain higher precision and faster calculating speed.