基于分形特征和阈值分析的图像边缘检测  

IMAGE EDGE DETECTION BASED ON FRACTAL FEATURE AND THRESHOLD ANALYSIS

作　　者：罗洪平 马洪兵[2] 王宏伟[3] Luo Hongping;Ma Hongbing;Wang Hongwei(School of Electrical Engineering,Xinjiang University,Urumqi 830046,Xinjiang,China;Department of Electronic Engineering,Tsinghua University,Beijing 100084,China;School of Control Science and Engineering,Dalian University of Technology,Dalian 110024,Liaoning,China)

机构地区：[1]新疆大学电气工程学院,新疆乌鲁木齐830046 [2]清华大学电子工程系,北京100084 [3]大连理工大学控制科学与工程学院,辽宁大连110024

出　　处：《计算机应用与软件》2024年第4期200-204,共5页Computer Applications and Software

基　　金：上海航天科技创新基金项目(SAST2019-048);自治区研究生科研创新项目(XJ2020G058)。

摘　　要：针对传统的边缘检测算法对噪声敏感且伪边缘较多,提出一种基于分形特征和阈值分析的图像边缘检测方法。该方法利用改进的毯覆盖算法计算出图像的分形特征,将图像的灰度分布映射到分形维数空间上,再根据提取的分形特征图进行阈值分析,获得高低两个阈值,将像素值分为非边缘、弱边缘和强边缘3类,再对弱边缘像素进一步加以判断。实验结果与其他算法相比较表明,该算法检测出来的图像边缘伪边缘和噪声最少。Traditional edge detection algorithms are sensitive to noise and have many false edges.This paper proposes an image edge detection method based on fractal features and threshold analysis.The improved blanket coverage algorithm was used to calculate the fractal features of the image,and the gray distribution of the image was mapped to the fractal dimension space.The threshold analysis was carried out according to the extracted fractal feature image to obtain the high and low thresholds.All the pixel values were divided into three categories:non-edge,weak-edge,and strong-edge,and the weak edge pixels were further judged.Compared with other algorithms,the experimental results show that the proposed algorithm can detect the image edge completely with the least false edge and noise.

关 键 词：边缘检测 分形特征 毯覆盖法 阈值分析 双阈值 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置] TP3[自动化与计算机技术—控制科学与工程]