提出了一种贝叶斯估计和尺度空间滤波相结合的滤波方法。信号的小波系数表现出很强的非高斯统计特性 ,其密度函数可以用推广的拉普拉斯先验分布建模。利用贝叶斯估计能够较好地提取出信号的小波系数 ,再由小波逆变换恢复信号。贝叶斯估计较好地保留了信号的边缘信息 ,但滤波信号尚不够平滑。尺度空间滤波在保留信号边缘的同时 ,还有很好的平滑作用。两者相结合有很好的滤波效果 A filtering method based on Bayesian estimation and scale spatial filtering is proposed. The wavelet coefficients of signals show strong non-Gaussian statistics, and their density functions can be modeled by a generalized Laplacian distribution. By using Bayesian estimation, the wavelet coefficients of the signal can be extracted well, and the signal is restored by wavelet inverse transform. Bayesian estimation better preserves the signal edge information, but the filtering signal is not smooth enough. Scale spatial filtering preserves the signal edge at the same time, there is a good smoothing effect. The combination of the two has a good filtering effect