结合模糊聚类与区域合并的彩色图像分割方法

提出了一种结合模糊聚类与区域合并的无监督彩色图像分割方法.首先,根据彩色图像建立对应的三维直方图,运用爬山法得到初始聚类中心和聚类数;然后,运用基于空间邻域像素的模糊均值聚类算法对图像在颜色空间进行聚类,得到初始分割结果;最后,应用提出的区域合并算法合并图像初始分割结果得到最终分割图像.仿真结果表明,算法的分割结果与人的主观视觉也有较好的一致性.

Color Image Segmentation Based on Fuzzy clustering and Region Merging

This paper proposed an unsupervised color image segmentation method combining fuzzy clustering and region merging. First, using Hill-climbing method to generate the initial centroids and the number of clusters adaptively according to the three dimensional histogram of the image. Second, the SFCM(A fuzzy c-means based spatial pixel and membership relationships) algorithm is employed to obtain the initial segmentation result in color space. Finally, a region merging algorithm is applied to get the final segmentation result. Experimental results indicate that the proposed method could generate reasonably better image partitioning.