一种用于多阈值图象自动分割的混合遗传算法

A New Mixed Genetic Algorithm for Multilevel Thresholding

模糊Ｃ－均值算法是一种比较有效的数据聚类方法．遗传算法则是一种借鉴生物界自然选择和自然遗传机制的高度并行、随机、自适应的搜索算法．该文有机地利用遗传算法与模糊Ｃ－均值算法，并考虑图象的二维灰度信息，提出了一种适用于多阈值图象自动分割的新方案．该方案能够快速正确地实现分割，且不需事先认定分割类数．

In this paper, a new approach for multilevel thresholding based on genetic algorithm and fuzzy C means algorithm has been proposed and implemented. Owing to the characteristics of genetic algorithm, the new approach has good robustness and global optimum. The approach has also been advanced to speed convergence and avoid local optimum. Experimental results are satisfied.