基于遗传算法的多小波自适应去噪方法研究  

MULTIWAVELET ADAPTIVE DENOISING METHOD BASED ON GENETIC ALGORITHM

作　　者：章琳[1] 方志军[2] 汪胜前[1] 杨凡[1] 刘国栋[1] 

机构地区：[1]江西科技师范学院光电子与通信重点实验室,江西南昌330013 [2]江西财经大学信息管理学院,江西南昌330013

出　　处：《红外与毫米波学报》2009年第1期77-80,共4页Journal of Infrared and Millimeter Waves

基　　金：国家自然科学基金资助项目(60662003,60462003)

摘　　要：针对噪声在多小波分解后的尺度性以及图像本身的特性,提出了一种基于遗传算法的多小波自适应去噪算法,该方法能通过遗传算法自适应地寻求去噪后图像的最小均方误差.实验结果表明,该算法优于传统算法,不仅能有效滤除图像的噪声,而且能较好地保留图像的边缘信息,具有更加理想的去噪效果.Aiming at the scaling of the noise after muhiwavelet decomposition and the characteristic of image, a multiwavelet adaptive denoising method was proposed based on genetic algorithm. This method can adaptivly look for the least RMSE of the denoised image according to genetic algorithm. The experimental results show that this algorithm is superior to the traditional methods. It can not only remove the noise of image, hut remain the better edge of image, and this method has more perfect denoising effects.

关 键 词：图像去噪 自适应阈值 遗传算法 图像特性 

分 类 号：TN911.73[电子电信—通信与信息系统]