高斯变邻域差分的灰度图像增强算法  

Gray image enhancement technology based on Gauss variable neighborhood differential evolution algorithm

作　　者：海洁[1] 吴海燕[1] 罗中剑[1] 

机构地区：[1]郑州大学西亚斯国际学院电子信息工程学院,郑州451150

出　　处：《激光杂志》2015年第1期57-61,65,共6页Laser Journal

摘　　要：灰度图像增强问题可以转化为一个目标优化问题,但是该目标函数实质是一个多峰值的优化问题,对此提出一种高斯变邻域差分进化算法用于灰度图像的增强。首先,利用高斯变异方式具有快速收敛特性以及变邻域方式在种群多样性保持方面的优势,对差分进化算法的变异进行改进,目的是平衡算法在快速收敛和保持种群多样性方面的能力,提高算法的整体效能。然后利用该算法对灰度图像增强问题进行研究,通过与基于标准DE,PSO及均衡直方图灰度图像增强效果进行对比,高斯变邻域差分灰度图像增强算法能够更有效的对灰度图像进行增强。The gray level image enhancement problem can be formulated to a optimization problem, the objective function is an optimization problem of multi peak, so, differential evolution algorithm based on the gauss and variable neighborhood mutation is used to enhance the image. Firstly, with the properties of the Gauss mutation fast conver-gence and variable neighborhood maintaining the advantages in the diversity of the population, the mutation of differen-tial evolution algorithm was improved, the objective is to balance algorithm in fast convergence and the ability to keep the population diversity, improve the overall efficiency of the algorithm. Then the algorithm of gray image enhancement problem is studied, and based on the standard DE, PSO and histogram equalization image enhancement effect was compared, the result shows that gauss variable neighborhood difference gray image enhancement algorithm can be more effective to enhance the gray image.

关 键 词：高斯变邻域 差分进化 灰度图像增强 

分 类 号：TN911.73[电子电信—通信与信息系统]