最优进化图像阈值分割算法

Optimal Evolution Algorithm for Image Thresholding

针对图像阈值分割问题,根据遗传算法理论提出最优进化图像阈值分割算法.将图像中每个像素点看作一个染色体,阈值看作进化方向,假设最优进化方向存在,建立进化方向更新模型;然后定义了染色体编码规则,通过简单随机采样进行种群初始化,重新定义了适值函数和选择机制,在适当的交叉率和变异率下得到最优阈值;同时分析了假设和模型的合理性.实验结果表明,文中的假设和进化方向更新模型合理,该算法是稳定、有效的图像阈值分割算法.

An optimal evolution algorithm for image thresholding is proposed based on theories of genetic algorithm. Pixel and threshold are regarded as chromosome and evolution direction, respectively. Assuming the optimal evolution direction exists, the updating model of evolution direction is established. Then by defining the chromosomes＇ coding rules, initializing the group by simple-random-sampling, and redefining the fitness function and the selection mechanism, the optimal threshold is obtained under the proper crossover rate and mutation rate. The rationalities of the assumption and the updating model have been analyzed in this paper. The experimental results show that the assumption and the updating model are proper; the proposed algorithm is robust and effective.