基于遗传模糊C-均值聚类算法的图像分割

将遗传算法（GA）与模糊C－均值聚类算法（FCM算法）相结合，并运用于图像分割，以期解决标准FCM算法在图像分割中运算速度慢和对初始值依赖大的两大缺陷。首先对模糊聚类中心进行编码，然后依据FCM算法的目标函数建立适应度函数，在适当的交叉率和变异率下，最终实现了基于遗传模糊C－均值算法的图像分割。考虑在一维图像分割特征向量情况下，通过引入直方图统计特性，实现了遗传模糊C－均值算法的快速运算，最后，运用真实的磨粒图像对算法进行了详细验证，并与标准FCM算法进行了对比，分割实验表明了本方法比标准FCM算法具有更快的计算速度和更好的鲁棒性。

Image Segmentation Based on the Genetic Fuzzy C-Mean Algorithm

In this paper, the Fuzzy C-Mean algorithm (FCM algorithm) and genetic algorithm (GA) are combined to overcome two shortcomings, namely the low computation speed of standard FCM algorithm in image segmentation and its over-dependence on initial values. Firstly, the fuzzy cluster center is coded; then the fitness function is established according to the object function in FCM algorithm, and under proper crossover rate and mutation rate, the image segmentation based on the genetic FCM algorithm is realized. Meanwhile, by taking into account 1-D image segmentation character vector and introducing the histogram, the speed of the genetic FCM algorithm is increased greatly. Finally, the genetic FCM algorithm is tested and compared with standard FCM algorithm in detail by using real wear particle images. Segmentation examples as shown in Figs.1 throngh 4 show that our method has faster computation speed and better robustness than standard FCM algorithm.