改进的小波变换阈值去噪方法

An Improved Thresholding Method in Wavelet Transform Domain for De-noising

针对Donoho的阈值去噪方法有时去噪效果不能令人满意的问题,提出了一种二次小波变换阈值去噪的方法——对小波变换各个尺度上的小波系数都进行二次小波变换,并进行去噪处理后重构小波系数,再由重构的小波系数重构原信号,此时原信号中已去除大部分随机噪声,此后再对重构后的原信号进行小波域阈值去噪。仿真结果表明:该方法具有良好的去噪效果,而且对信号的主要细节保留较好。

Donoho's de-noising method by threshold filter in wavelet domain can remove random noises,but its de-noising effect is not always good.A new improved method of removing random noises is proposed in this paper,which combines quadratic wavelet transform with threshold filter.The simulation result shows that the effect of the method is good for removing the random noises,and also can remain the signal's main details well.