基于反对称双正交小波分解系数的模极大值的信号快速重构
doi: 10.3724/SP.J.1146.2005.01480
Fast Signal Reconstruction from Modulus Maxima Based on Anti-Symmetrical Biorthogonal Wavelets
-
摘要: 因为反对称双正交小波具有近似微分算子性质,该文用它对信号作无下采样(Un-decimated)的小波变换,以此获得信号在各尺度下与信号的边缘点相对应的局部模极大值。在只保留这些局部模极大值(即边缘)点信息的条件下,利用高斯函数直接拟合各级小波系数,然后再作小波逆变换就可恢复原信号。实验表明,与现有的采用二进小波变换的方法比较,该方法更加简便快捷。Abstract: The signal is done un-decimated wavelet transform using Anti-Symmetrical Biorthogonal Wavelets (ASBW). Because of the ASBWs approximate derivative property, the local maxima of the wavelet decomposition coefficients can be obtained, which is the corresponding edge of the signal at every scale. Only keeping the local maxima, the Gaussian function is used to fit the wavelet coefficients directly, then inverse wavelet translation is used to reconstruct signal. Numerical results show that the reconstruction algorithm is much faster than the dyadic wavelet translation.
-
Mallat S and Zhong S. Characterization of Signals from Multiscale Edges[J]. IEEE Trans. on PAMI, 1992, 14(7): 710-732.[2]蔡汉添, 宋勇. 关于子波变换局部极大值信号重构的交替投 影算法[J]. 数据采集与处理, 1998, 13(2): 112-116.[3]Liew A W C and Law N F. Reconstruction from 2-D wavelet transform modulus maxima using projection[J].IEE Proceedings on Vision, Image and Signal Processing.2000, 147(2):176-184[4]陈德智,唐磊. 由小波变换的模极大值快速重构信号[J]. 电子学报, 1998, 26(9): 82-85.[5]廖国勇,徐长友. 一种快速重构信号的方法[J]. 华中科技大学学报, 2002, 30(8): 98-100.[6]彭进业,俞卞章,王大凯. 多尺度对称变换及其应用于定位人脸特征点[J].电子学报, 2002, 30(3): 363-366.[7]Vetterli M and Herley C. Wavelets and filter banks: Theory and design [J]. IEEE Trans. on Signal Process. 1992, 40(9): 2207-2232.[8]王大凯,彭进业. 小波分析及其在信号处理中的应用. 北京:电子工业出版社, 2006.1, 第7章第2节, 第6章第1节.
计量
- 文章访问数: 3537
- HTML全文浏览量: 78
- PDF下载量: 889
- 被引次数: 0