由小波变换的模极大值快速重构信号

本文从多尺度分析的思想出发，提出一种由信号小波变换的模极大值快速重构信号的方法，在不同尺度上，以阶跃函数在该尺度上的小波变换作为基函数，利用模极大值拟合信号的小波变换，然后进行小波反演得到重构信号。这种快速算法具有明确而直观的物理意义，实现简单方便。文中还给出了对重构信号的误差修正方法，数值方法的结果表明，这种快速重构算法对不同类型的信号同样有效。

由小波变换的模极大值快速重构信号

Based on the multiscale idea,a fast algorithm to reconstruct the signal from its wavelet transform modulus maxima is built, in which the wavelet transforms of a step function are used as bases to reconstruct the wavelet transform of the sigul according to the medulus maxima on every scale, and then all of these components are used to reconstruct the signal by inverse wavelet transform. This methed is of explicit physical meaning and can be implemented conveniently. The methed of error medification is also given in this paper. Numerical results show that the fast reconstruction algorithm is effective for different types of signals.