摘要
边缘检测是提取图像特征的重要手段,本文提出一种阈值自动设定的双阈值小波变换边缘检测方法。小波变换具有良好的时频局部化特性和多尺度分析能力,获得的边缘信息丰富,采用二维静态小波变换算法,计算出局部模极大值点,但其中除了边缘点外,还混有噪声信号,本文根据边缘与噪声的特征自动计算出阈值,实现了噪声与边缘的分离,强边缘与弱边缘的分离,并通过实例验证了算法的有效性。
Edge detection is an important method to extract the image' s characteristics. The paper puts forward a thresholding auto-matic setting method for edge detection based on wavelet.Wavelet transform has nice time-frequency localization characteristic and multiscale analysis ability, so it can obtain abundance edge information. Two-dimensional stationary wavelet transform is adopted to compute local modulus maximum. It compute the thresholding automatically according to the characteristics of edge and noise. Large number experiments show that this method can separate edge form noise, and strong edge from weak edge.
出处
《微计算机信息》
北大核心
2007年第3期286-287,308,共3页
Control & Automation
基金
华东交通大学校立科研基金资助(06ZKXX02)
关键词
边缘检测
静态小波变换
局部模极大值
阈值
edge detection,stationary wavelet transform,local modulus maximum,thresholding