基于最优匹配小波的回波与图像消噪

Denoising for ultrasound echo and image based on optimal match wavelet

研究了用于消噪的小波与信号、图像的匹配原则,在采用时域波形逼近与频谱匹配原则的基础上,提出了最优匹配小波的设计方法。首先构造了两种小波滤波器组,其次给出了最佳匹配小波的设计优化算法,最后分别探讨了应用最优匹配小波对超声回波和图像消噪的方法。仿真结果表明,该方法性能优于常规的小波消噪,能较明显提高信号与图像的信噪比。

The choices of wavelet bases and threshold are two key problems in using wavelet transform to process signal and image. Aimed at the first problem, an algorithm for designing optimal match wavelet is proposed. Two kinds of wavelet filter banks based on this algorithm are constructed. The optimal algorithm of the match wavelet is presented. Criterions of wavelet matching signal or image are also put forward. It is shown that the construction of match wavelet is equivalent to the optimal design of perfect reconstruction filter banks which match the signal or image. The method of denoising is studied with optimized match wavelet transform for ultrasonic echo and image. The simulation results show that this method is suprior to normal wavelet method, it can increase the ratio of signal and noise. 2 tabs, 8 figs, 6 refs.