摘要
为了快速准确地确定图像的最佳分割阈值,提出了一种改进的遗传算法。该算法通过完善选择机制、引进父子竞争机制和使用二元变异算子进行变异操作,有效地解决了遗传算法的收敛速度慢和种群过早成熟的问题。实验结果表明:采用改进遗传算法对图像搜索阈值时能收敛至全局最优解,并且大大缩短寻找最优阈值的时间,取得良好的分割效果。
In order to rapidly and accurately determine the optimal threshold in image segmentation, an improved genetic algorithm is proposed. Through consummating the choice mechanism, introducing competing between parents and children and using dyadic mutation operator, it can overcome the problems of poor astringency and premature occurrence in Genetic Algorithm. The experiment results show that the segmentation method can greatly improve the speed of segmentation and can effectively overcome the premature convergence.
出处
《浙江理工大学学报（自然科学版）》
2008年第6期700-703,共4页
Journal of Zhejiang Sci-Tech University(Natural Sciences)
基金
国家基金(60773204)
浙江省自然科学基金(Y107124)
关键词
遗传算法
图像分割
父子竞争
二元变异算子
最大类间方差法
genetic algorithm
image segmentation
competing between parents and children
dyadic mutation operator~ Otsu
