【摘要】 传统的小波变换软阈值消噪方法是基于分布不均匀的信号和白噪声在小波变换下具有不同的传播特性 ,该方法比较简洁 ,并在一定程度上是有效的 .在这种方法的基础上 ,提出了一种估计小波系数的模平方处理改进方法 ,它可以促使信噪分离 .与原有的方法相比 ,采用改进方法明显地提高了信号消噪所得信噪比增益 .更多还原

【Abstract】 The traditional me thod for white noise reduction with wavelet soft-thresholding is based on the d ifferent propagation property between unevenly distributed signals and white noi ses under the wavelet transform. This method is concise and effective to some ex tent. An improved modulus square processing method to estimate the wavelet coeff icients is presented in this paper, which can accelerate the separation of the s ignal from the noise. Finally, an example is given to show that the improved met hod is superior to the primary one and it can get higher SNR gain for the recons tructed signal.更多还原