一种自适应FCM彩色图像分割算法*1

针对FCM彩色图像分割算法需要给定聚类数且受初始值影响较大的缺点,提出一种自适应FCM彩色图像分割算法。该方法首先引入主分量变换,通过直方图分析,在RGB彩色空间上自动确定聚类数目,其次引入2D直方图,分析对角线投影1D直方图,并采用最大最小距离法选取初始聚类中心。利用该算法分别对人工合成的和真实的彩色图像进行分割实验,实验结果表明：本文算法能够有效且自适应地分割彩色图像。

Based on the principal component 2D histogram analysis and adaptive FCM algorithm of color image segmentation method

In accordance with the need of FCM color image segmentation algorithm given cluster number and greatly influenced by the initial value of faults, puts forward an adaptive FCM color image segmentation algorithm. The algorithm firstly introduces principal component transformation, through the analysis of the histogram, automatically determine the clustering number in RGB color space, then the introduction of 2D histogram, 1D histogram analysis of diagonal projection, and maximum minimum distance method is used to select the initial cluster centers. Respectively using the algorithm of synthetic and real color image segmentation experiments, the experimental results show that this algorithm can effectively and adaptive color image segmentation.