Continuous wavelet transforms (CWT) have lots of applications in the field of signal processing due to its unique characteristics. Their realization, however, request considerable computation. Therefore, a fast algorithm is provided to combat this contradiction. This algorithm is realized by organizing some computing cells which are built by two filters f(n) and g(n). In this paper, some practicable methods are discussed to construct these filters. Moreover, the structure of the whole fast algorithm, along with a method to refine the scale interval of wavelet coefficients, is also presented.
Goupillaud P, Grossmann A, Morlet J. Cycle-octave and related transforms in seismic signal analysis.[2]Geoexploration. 1984, 23(1): 85-102.[3]杨福生.信号的时间-频率分析,北京:清华大学电机系.1994,第四章,39-56.[4]彭玉华.利用小波变换对目标的电磁场后向散射信号进行时频分析:[博士论文].西安西安交通大学,1994年.[5]Rioul O, Duhamel P. Fast algorithm for discrete and continuous wavelet transforms. IEEE Trans. on IT, 1992, IT-38(2):569-586.[6]Shensa M J. The discrete wavelet transform:Wedding the a Trous and Mallat algorithm. IEEE Trans. on SP, 1992, SP-40(10): 2464-2482.[7]刘贵忠.Shannon多分辨分析.信息与控制,增刊,1995,423-434.[8]刘贵忠,冯轶岭,宗涛.样条多分辨分析.电子学报,1996,24(7):72-77.[9]Daubechies I. Orthonormal bases of compactly supported wavelet. Comm. on Pure and Applied Mathematics, 1986, 41(7): 909-996.