摘要
提出了一种改进的小波变换快速算法.通过对小波滤波器系数特点的分析,发现M allat算法的速度可以进一步提高.在小波分解与重构中可以成对地计算,即在分解中一次计算一个低频系数和一个高频系数,而在重构中则一次计算相邻两个恢复值.在每一对值的计算过程中,后一个值的计算可以利用前一个值的计算结果,从而减少乘法和加法的次数,达到提高速度的目的.文中给出了M allat算法和改进算法的实验对比数据.
An improved fast algorithm based on wavelet transform is provided. The analysis on wavelet filter coefficients shows that the speed of Mallat algorithm can be improved. A low-frequency coefficient and a highfrequency coefficient can be calculated simultaneously in the decomposition and reconstruction of wavelets, and two adjacent recovered values can also be calculated simultaneously in reconstruction. During the calculation of a pair of values, the later value can be calculated based on the result of the former value. Thus the times of multiplication and addition are decreased and the speed is improved. The comparative data between Mallat algorithm and the improved fast algorithm are provided, which illustrate the advantages of the improved algorithm.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2009年第9期214-217,共4页
Journal of Harbin Institute of Technology
基金
国家高技术博士点资助项目(20030145017)
关键词
小波变换
滤波
分解
重构
wavelet transform
filter
decomposition
reconstruction