多种群遗传算法在图像处理中的应用研究  

Application of Multi Population Genetic Algorithm in Image Processing

作　　者：刘拓 LIU Tuo(Wuhan University of Science and Technology College of Information Science and Engineering,Wuhan Hubei 430081)

机构地区：[1]武汉科技大学信息科学与工程学院,湖北武汉430081

出　　处：《软件》2023年第8期143-146,共4页Software

摘　　要：传统图像增强方式存在一系列缺陷,基于此有针对性地提出以多种群遗传算法为基础的新图像增强算法。首先,将非完全Beta函数图像增强过程参数编码成为遗传算法中的个体,然后根据适应度函数,模拟自然界生物进化过程对参数进行寻优,这是一种标准的遗传算法,因为优化过程中不依赖于梯度,使其具有很强的鲁棒性和全局搜索能力,但存在一个不可忽视的问题,即过早收敛,主要表现为种群中的所有个体都趋于同一状态并结束进化,导致最终算法无法获得相对满意的结果。为了克服上述问题,多种群遗传算法引入移民算子和精华种群,实验表明它具有很好的稳定性和较快的收敛性。In order to solve the defects oftraditional image enhancement methods,this paper proposes an image enhancement algorithm based on multi population genetic algorithm.Firstly,the image enhancement process parameters of the incomplete Beta function are encoded as individuals in the genetic algorithm,secondly,according to the fitness function,the natural biological evolution process is simulated to optimize the parameters,this is a standard genetic algorithm,because the optimization does not depend on gradient,it has strong robustness and global search ability,but there is a problem that cannot be ignored,namely,premature convergence,the main performance is that all individuals in the population tend to the same state and stop evolution,and the algorithm cannot give a satisfactory solution finally.In order to overcome the above problems,the multi population genetic algorithm introduces the migration operator and the essence population,experiments show that it has good stability and fast convergence.

关 键 词：图像增强 多种群遗传算法 非完全Beta函数 标准遗传算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]