航空学报 > 2009, Vol. 30 Issue (9): 1697-1704

参数化小波构造研究及其在边缘检测中的应用

祝强1,李言俊2,徐瑾3,张建华2   

  1. 1 西安工业大学 机电工程学院 2 西北工业大学 航天学院 3 武汉理工大学 机电工程学院
  • 收稿日期:2008-05-25 修回日期:2008-12-22 出版日期:2009-09-25 发布日期:2009-09-25
  • 通讯作者: 祝强

Research on Parameterized Design of Wavelet and Its Application in Edge Detection

Zhu Qiang1, Li Yanjun2, Xu Jin3, Zhang Jianhua2   

  1. 1 Mechanical & Electrical Engineering Institute, Xi’an Technological University 2 School of Astronautics, Northwestern Polytechnical University 3 Mechanical & Electrical Engineering Institute, Wuhan University of Technology
  • Received:2008-05-25 Revised:2008-12-22 Online:2009-09-25 Published:2009-09-25
  • Contact: Zhu Qiang

摘要: 为了解决双正交小波构造过程中存在的局限性和不灵活性,以小波滤波器代数构造理论为基础,提出一种支集定长条件下双正交小波滤波器组的参数化构造方法,推导出高通分解滤波器支集长度为奇数或偶数情况下双正交小波滤波器组的构造限定条件,给出具体的参数化构造实例,并分析了参数变化对构造小波特性的影响。从小波特性出发分析了用于图像边缘检测的双正交小波的构造必要条件,结合参数化小波构造方法构建了用于图像边缘检测的小波滤波器组。仿真实例验证了双正交小波滤波器参数化构造方法的灵活性和可行性以及构造小波对图像边缘检测的有效性。

关键词: 滤波器组, 边缘检测, 小波构造, 小波特性, 参数化

Abstract: To solve the inflexibility and the limitation of the biorthogonal wavelet construction, a filter banks parameterized design in case of a fixed support length is proposed which is based on filter banks algebraical construction method. The limiting conditions of biorthogonal wavelet filter banks construction are derived for cases when the filter length of high pass decomposing filter banks is odd or even. According to the new construction method, some filter banks parameterized design examples are given. And the effects on biorthogonal wavelets’ characteristics is discovered when the parameter is changed. By researching the wavelet characteristics, the necessary construction conditions of biorthogonal wavelet which is used in edge detection are analyzed. And the wavelet filter banks used in image edge detection are built in term of parameterized design method. The simulation results show that the filter banks parameterized design of biorthogonal wavelet is feasible and flexible. The built wavelets are effective in image edge detection.

Key words: filter banks, edge detection, wavelet construction, wavelet characteristic, parameterized

中图分类号: