The present method for noise reduction with wavelet transform is based on the thresholding method and threshold selection will depend on the estimation of the noise standard deviation.
目前的小波去噪绝大多数都是基于阈值决策的方法，而阈值选择往往都是建立在对信号小波变换系数的方差估计上。
This paper achieves noise reduction by using wavelet packet transform, semisoft shrinkage function, Donoho standard deviation estimate and threshold obtain based on statistics.
计算结果表明：变形统计值与变形监测值吻合较好，统计复相关系数较大，估计标准误差较小。
应用推荐