基于小波和模糊集的多尺度边缘检测算法  

Multi-scale Edge Detection Based on Wavelet and Fuzzy Sets

作　　者：李惠光[1] 尹玉[1] 李旭峰[1] 张磊[1] 

机构地区：[1]燕山大学电气工程学院,河北秦皇岛066004

出　　处：《控制工程》2007年第S3期23-25,29,共4页Control Engineering of China

摘　　要：针对传统边缘检测算子对噪声敏感问题和使用固定阈值小波变换造成的强弱边缘,提出了一种基于B样条小波变换和广义模糊算子(GFO)的边缘检测方法。首先对图像进行多尺度分解,并应用广义模糊算子对各尺度子图增强,再结合自适应阈值选取的方法,在每种尺度下分别提取图像边缘;而后利用边缘信息的多尺度特性,融合多尺度边缘得到了单像素宽边缘。通过计算机仿真对该方法进行验证,实验结果表明该方法不仅能准确检测出图像边缘,而且能有效地抑制噪声,效果优于传统的边缘检测算法。To the problem that traditional edge detectors are sensitive to noise and blur edges caused by fixed threshold wavelet transform,an edge detection algorithm based on B-spline wavelet and general fuzzy operator is proposed.First,image is decomposed at multi-scale,and en- hanced by general fuzzy operator(GFO),then combined with adaptive thresholds to extract image edge at different scales.The multi-scale edge with only a pixel width is obtained by fusing different scale edges with the multi-scale characteristic of edge information.Computer simulation re- suits show that the proposed method can not only detect image edge precisely,but also effectively restrain noise.Comparing with several tradi- tional algorithms,the experiment results show the feasibility and effectiveness.

关 键 词：小波变换 多尺度分析 广义模糊算子 图像增强 自适应阈值 边缘检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]