基于区间值直觉模糊集的图像边缘检测算法  

Image edge detection algorithm based on interval valued intuitionistic fuzzy sets

作　　者：陆盈 Lu Ying(School of computer and information engineering,Nantong University of Technology,Nantong 226002,China)

机构地区：[1]南通理工学院计算机与信息工程学院,南通226002

出　　处：《电子测量与仪器学报》2018年第12期99-107,共9页Journal of Electronic Measurement and Instrumentation

基　　金：国家自然科学基金(61572239);江苏省自然科学基金(BK20170113)资助项目

摘　　要：针对当前边缘检测算法中边缘定位精度低、噪声抑制能力较弱等问题,提出了一种基于区间值直觉模糊集(IVIFS)的图像边缘检测算法。首先,对于给定的灰度图像,将其转换为模糊图像;其次,利用T范数和T-三角余模,构造区间值模糊集(IVFS)的上、下极值,并将模糊图像划分为区间值模糊图像;然后,根据隶属度和非隶属度区间函数,通过区间值直觉模糊生成器来构造IVIFS图像。再根据边缘像素与相邻像素强度关系,得到直觉模糊边缘信息,利用去模糊机制获取模糊边缘。最后,通过线性方程的模糊指数优化,生成图像边缘。实验表明,与当前边缘检测方案比较,所提技术得到更多的边缘细节,完整性与清晰度更好,边缘定位精度高,具有较强的噪声抑制力,适用性强。Aiming at the problems of low edge location accuracy and weak noise suppression ability in the current edge detection algorithm,an image edge detection scheme based on interval-valued intuitionistic fuzzy sets ( IVIFS) was proposed. Firstly,for a given grayscale image,it was converted to blurred images. Secondly,the fuzzy images were divided into interval valued fuzzy images by using the T-norm and the T-triangular residual model to construct the upper and lower extremes of the interval-valued fuzzy sets ( IVFS) . Then,interval valued intuitionistic fuzzy generators were introduced to construct IVIFS images according to membership and nonmembership interval functions. Then the intuitionistic fuzzy edge image was obtained based on the intensity relationship between edge pixels and adjacent pixels,and fuzzy edge was generated through deblurring. Finally,the edge of image was generated by fuzzy exponential optimization of linear equation. The experiment shows that compared with the current edge detection scheme,this scheme has more edge details,better integrity and sharpness,as well as higher precision of edge location,stronger noise suppression and wider applicability.

关 键 词：图像边缘检测 区间值直觉模糊集 区间值模糊集 模糊边缘 边缘像素 去模糊机制 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]