J4

• 数学 •    下一篇

二维非张量积小波用于图像的边缘检测

李瑛, 邹溪   

  1. 吉林大学数学研究所, 长春 130012
  • 收稿日期:2004-01-04 修回日期:1900-01-01 出版日期:2004-10-26 发布日期:2004-10-26
  • 通讯作者: 李 瑛

Image edge detection by means of bivariate non-tensorproduct wavelet

LI Ying, ZOU Xi   

  1. Institute of Mathematics, Jilin University, Changchun 130012, China
  • Received:2004-01-04 Revised:1900-01-01 Online:2004-10-26 Published:2004-10-26
  • Contact: LI Ying

摘要: 构造了二元非张量积紧支集对称的连续预小波, 分析了所构造的紧支集、 对称的、 非张量积预小波适合于图像边缘检测的原因, 针对所构造的非张量积小波的特点, 提出了有效的边缘检测算法, 其中噪音与边缘信息的区分非常重 要. 用所构造的非张量积预小波对不同的图像进行了边缘检测, 所提出的边缘检测方法明 显优于Sobel边缘检测算子, 与Canny最优边缘检测算子的效果与复杂性均持平, 并且对纹 理性强的图像(如指纹)边缘检测效果优于其他算法.

关键词: 二元预小波, 紧支集, 非张量积, 边缘检测

Abstract: We constructed bivariate non-tensor compactly supported prewavelet with symmetrical property, and discussed why our constructed prewavelets are adapted to image edge detection, and put forward an effective edge detection algorithm of non-tensor wavelets according to the feature of our constructed non-tensor wavelets, in which the key is how to distinguish the noise from the image edge information. For different images, we detected edge by our method, compared the detechion with Sobel edge detection operator and Canny edge detection. From the examples, the performance of our method is obviously better than Sobel edge detection and is almost equal to Canny edge detection, and for the image with visible texture such as the fingerprint image, our method is a little b etter than other methods.

Key words: bivariate prewavelet, compactly supported, non-tensor product, edge etection

中图分类号: 

  • O241