摘要
文章利用小波变换和多分辨率分析的性质,从多尺度角度对图像边缘检测算法进行分析,总结出小波变换模极大值多尺度边缘检测算法。通过对标准图像Lena进行小波变换模极大值多尺度边缘检测结果发现,其比小波变换模极大值边缘检测算法和Canny算法在部分边缘检测中得到了更多的细节信息,使图像变得更真实。
This paper taked advantage of the property of wavelet transform and muhiscale resolution analysis and maked a analysis of edge detection of image from muhiscale point of view and concludes multiscale edge detection algorithms which bases on wavelet transform modulus maxima. Through the result of the standard image of Lena' s edge detection, we found that muhiscale edge detection algorithms can get more details than edge detection algorithms and Canny algorithms and make the mage more real.
出处
《四川理工学院学报(自然科学版)》
CAS
2009年第6期87-91,共5页
Journal of Sichuan University of Science & Engineering(Natural Science Edition)