基于小波熵和相关性的高分辨率阈值去噪方法

摘要

提出了一种基于小波熵和相关性相结合的高分辨率小波阈值去噪方法.首先利用小波变换中各尺度间有效信息和噪声的相关性不同的特性,对小波分解后的各尺度的高频小波系数进行相关处理,确定出有效信息的位置,并将其置零,经过相关处理后的高频小波系数认为是由噪声引起的.将相关处理后的高频小波系数分成若干区间,计算各区间的小波熵,将小波熵最大区间的高频小波系数的平均值作为噪声标准差,计算各尺度的阈值;采用软阈值处理,最后重构得到去噪后的信号.该算法实现了各尺度阈值的自适应选取,提高了信噪比.仿真验证了该算法的有效性.%A high-resolution threshold denoising method combined wavelet entropy with correlation is proposed.The correlation of the effective signal and noise is different.The effective signal location is determined by the correlation processing of high-frequency wavelet coefficient of each level,and the high-frequency wavelet coefficient of the effective signal is zero.The retained high-frequency wavelet coefficient is caused by noises.The high-frequency wavelet coefficient of each level is divided into several small zones,and the interval wavelet entropy is calculated,with the mean value of high-frequency wavelet coefficients in the wavelet entropy maxima interval as noise standard deviation.Threshold value of each level is calculated.The retained coefficient is processed by soft-threshold and the denoised signals are reconstructed.The improved method realizes adaptive selection of threshold values of each level and the improved signal-to-noise radio simulation verifies the effectiveness of the improved method.