基于模糊C均值聚类的遥感图像分割方法

在遥感图像分割领域,模糊C均值聚类算法得到了广泛的应用。但存在计算量大、易受噪声干扰等缺点。针对以上缺点对快速模糊C均值聚类算法进行了改进。首先利用一维灰度直方图进行快速模糊C均值聚类降低计算量;然后在此基础上根据像素的邻域特性构造新的隶属度函数;最后根据新的隶属度函数对每个像素进行分类。实验结果表明,该算法能快速有效地分割图像,并具有较强的抗噪能力。

Remote sensing image segmentation based on fuzzy C-means clustering algorithm

FCM clustering algorithm is widely applied to automated image segmentation.But standard FCM algorithm has many problems,such as great amount of calculation and easy noise incernecine.This paper proposes a modified fast FCM algorithm for image segmentation.With the modified algorithm,images can be mapped to gray-scale histogram space from pixel space,on the basis of which membership function can be improved by the full use of pixel's neighborhood feature.The experiment shows that new algorithm is effective i...