三维最大Renyi熵的灰度图像阈值分割算法  

Thresholding algorithm based on three-dimensional Renyi's entropy

作　　者：魏巍[1,2] 申铉京[1,2] 千庆姬[3] 李端[4] 

机构地区：[1]吉林大学计算机科学与技术学院,长春130012 [2]吉林大学符号计算与知识工程教育部重点实验室,长春130012 [3]吉林大学物理学院,长春130015 [4]中国联通佳木斯分公司,黑龙江佳木斯154002

出　　处：《吉林大学学报（工学版）》2011年第4期1083-1088,共6页Journal of Jilin University:Engineering and Technology Edition

基　　金：国家自然科学基金项目(60773098);吉林省自然科学基金项目(201115025)

摘　　要：针对现阶段的阈值分割算法(如最小误差法、Otsu法、Renyi熵法等)在灰度、邻域均值或中值构成的二维或三维直方图上进行分割而造成的误分和低抗噪性等问题,提出了一种基于灰度、邻域均值和邻域加权中值三维直方图的最大Renyi熵阈值分割算法。相比目前多种阈值分割算法,新加入的邻域加权中值既能很好地过滤噪声,又能保留一定的图像边缘细节。经过对多幅图像的实验表明,该方法无论在精度上还是抗噪性上都有明显提升。The current thresholding methods are mostly based on 2D or 3D histograms of grayscale,neighbor average grayscale and neighbor median grayscale.These methods include 2D minimum error method,2D Renyi＇s entropy method,3D Otsu method etc.The segmentation precisions and noise immunities of these methods are not satisfactory.A novel thresholding algorithm based on 3D Renyi＇s entropy is proposed,the dimension added is the weighted neighbor median grayscale.Compared with other existing methods,it can not only satisfactorily filter the noise,but also reserve the edge details to a certain extent.Experiment results on many images indicate that this algorithm is able to promote the precision and noise immunity effectively.

关 键 词：计算机应用 阈值分割 RENYI熵 三维直方图 邻域加权中值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]