嵌入灰度级信息的粗糙模糊熵阈值分割法  

Gray level information embedded rough-fuzzy entropy thresholding method

作　　者：张干[1] 吴成茂[1] 

机构地区：[1]西安邮电大学电子工程学院,陕西西安710121

出　　处：《计算机工程与设计》2014年第2期541-545,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61073106;90607008);陕西省教育厅自然科学基金项目(2013JK1129)

摘　　要：针对现有基于信息熵、最大模糊熵、最大粗糙熵和粗糙模糊熵等分割方法在去噪以及保留图像细节信息等方面存在的不足,提出一种灰度级大小信息融合粗糙模糊熵的阈值分割法。将传统灰度图像采用粗糙模糊集描述所对应的不确定粗糙模糊熵与图像灰度级大小相融合,获得一种参数化的变权粗糙模糊熵的图像阈值化分割新准则。实验结果表明,该方法有较好的抗噪性和保留图像细节信息能力,在参数选取特定值时,可将传统粗糙模糊熵阈值法视为该方法的特例。To solve the problem that the existing thresholding method like information entropy, maximum fuzzy entropy and rough-fuzzy entropy are all have shortcomings in denoising and preserving image details, an gray level information embedded rough-fuzzy entropy image thresholding method is presented. Rough-fuzzy entropy defined by the rough-fuzzy sets is integrated into gray level information. A parametric weights rough-fuzzy entropy image thresholding method is obtained. The experimental results show that the method has good ability to preserve image details and to denoise. When a specific value of the parameter is selected, the traditional rough-fuzzy entropy thresholding method becomes a special case of this method.

关 键 词：信息熵 粗糙模糊 灰度级权重 去噪 阈值分割 

分 类 号：TN911.73[电子电信—通信与信息系统]