改进免疫遗传算法用于图像阈值分割  

Improved Immune Genetic Algorithm for Image Threshold Segmentation

作　　者：魏衍君[1] 陈哲[1] 郑青碧[1] 

机构地区：[1]商丘职业技术学院计算机系,河南商丘476000

出　　处：《河南科学》2012年第1期81-84,共4页Henan Science

基　　金：商丘市科技攻关计划项目(20102009)

摘　　要：在图像阈值分割中,基于遗传算法的分割方法存在着运行速度慢、易形成未成熟收敛等缺点.针对这一问题对其进行了改进.改进的免疫遗传算法在免疫算子中引入疫苗接种机制,极大地提高了收敛效率,对交叉概率和变异概率进行了改进,避免了局部收敛,以保证改进算法能收敛到全局最优值.实验结果表明,改进的免疫遗传算法比传统的算法提高了运行效率,解决了全局搜索不收敛和局部搜索不到最小值的问题,并具有更好的收敛稳定性.Among image threshold segmentations, there exist shortcomings, such as low speed and easily forming premature convergences, in segmentation method based on genetic algorithm. So a new improved immune genetic algorithm is proposed. The improved algorithm adds vaccination mechanism based on immune adjustment operator, so it greatly improves the convergence efficiency, crossover probability and mutation probability to avoid local convergence, ard to ensure the improved algorithm to converge the global optimum value. The experiment results show that the improved immune genetic algorithm can increase the operational efficiency, solve the convergence problem of the global search, local search and convergence stability.

关 键 词：图像阈值分割 免疫遗传算法 收敛 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]