基于改进遗传算法的最佳阈值分割方法及其性能评价  

Maximum entropy threshold segmentation method based on improved genetic algorithm and its performance evaluation

作　　者：隋然 潘点飞 

机构地区：[1]全军后勤信息中心,北京100842 [2]航天员科研训练中心,北京100094

出　　处：《微型机与应用》2015年第14期45-47,共3页Microcomputer & Its Applications

摘　　要：针对常规二维最佳熵法计算复杂,运行时间长,收敛性差等不足,提出基于改进遗传算法的二维最佳熵阈值分割方法。通过对选择、交叉、变异等因子的优化设计,使阈值搜索的鲁棒性与收敛性有了很大改善,并对图像的分割效果进行评价。分析与仿真结果表明,改进算法在大大减少阈值搜索时间的同时,保持了良好的分割性能。Aiming at the problems of the classical algorithms for solving two-dimensional maximum entropy , such as heavy computational complexity, long running time and poor convergence reliability, a novel method based on improved genetic algorithm is proposed to solve the maximum entropy. The improved selection operator, crossover operator and mutation operator are used to search threshold, which have improved dramatically on the robustness and convergence reliability. Then the image segmentation effect is evaluated. The results show that the improved method greatly saves the search time of threshold, as well as maintains good image segment effect.

关 键 词：二维最佳熵 遗传算法 图像分割 评价指标 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]