对称小波变换用于非平稳信号检测和去噪

根据信号与噪声在小波变换下表现出截然不同的性质,提出了一种基于对称小波多层系数乘积的信号去噪算法,该算法首先对信号进行多尺度小波变换,通过相邻尺度小波系数乘积提取小波变换模极大值,归一化和去除噪声小波系数;再利用归一的模极大值与相邻低层小波系数再次相乘作为对应层次的小波系数,逆变换得到去噪后的信号。实验仿真表明:该算法在有效去除噪声的同时,还能突出信号中的突变点。

On Symmetry Wavelet Transformation Applied to Examine Non-stable Signal and Remove Noise

In view of different properties of signal and noise in wavelet transformation,a new signal de-noising algorithm based on multiscale product in wavelet domain is presented.The initial signal is acted by the multiscale dyadic wavelet transform.Then the wavelet coefficients of the wavelet transform modulus maximum are extracted where the noising wavelet coefficients are removed by the use of adjacent scale wavelet coefficient product.The de-nosing signal is obtained through inverse wavelet transform using wavelet coefficients of modulus maximum.Simulation experiments show that the algorithm can remove the noise and preserve signal's original features effectively.