基于邻域加权模糊C均值的遥感影像分割

遥感影像分割是遥感影像识别和理解的前提和基础,遥感影像分割的结果直接决定着后续像分析和理解的质量。针对经典模糊C均值(Fuzzy C-Means,FCM)聚类算法分割图像时存在的不足,即该算法只考虑图像中像素自身的灰度值信息而没有考虑其邻域内的像素空间信息从而对噪声比较敏感,提出了一种邻域加权FCM(Neighboring Weighted FCM,NW-FCM)的遥感影像分割算法,该算法中邻域窗口内各系数(权)的值是根据图像自身的特性而自适应确定的。通过分割合成图像和实际的高分辨率遥感影像的实验结果表明,相对于其他几种方法,所提的方法取得了更好的分割效果。

Segmentation of Remote Sensing Image Based on Neighboring Weighted Fuzzy C-Means

Segmentation of remote sensing image is the basis of remote sensing image recognition and understanding.Segmentation result of remote sensing image directly determines the quality of successive image analysis and understanding.For the shortcomings of the traditional FCM algorithm in the process of image segmentation,i.e.,it is sensitive to the noise due to the fact that it only accounts for the pixel’s value information,not takes the neighboring pixels’ spatial information into consideration,and a neighboring weighted FCM(NW-FCM) algorithm for remote sensing image segmentation was proposed.The values of the coefficients within neighboring window were adaptively determined by the characteristics of the image itself.Experimental results on synthetic image and real remote sensing images with high resolution show that the proposed algorithm achieves better performance compared with other algorithms.