基于小波熵自适应最佳分解层数确定算法

摘要

针对超声信号小波阈值去噪中最佳分解层数的选取问题，提出基于小波熵的自适应分解层数确定算法。该算法首先利用离散小波变换分解含噪信号，计算分解后信号子区间的小波熵，然后将细节系数与原始信号的熵之比和低频子带均方根误差相结合以确定最佳分解层数。最后利用信噪比( SNR)、均方根误差( RMSE)、峰值相对误差( REPV)和峰位置误差(EPP)四项指标对算法性能进行评估。仿真和实验的结果表明：该算法能自适应地确定最佳分解层数，在有效滤除含噪超声信号中的噪声、提高性噪比的同时，还能更有效地保留原始信号中的有用成分。%Aiming at the problem of how to select the optimal level of wavelet for ultrasonic signal denoising, an improved wavelet entyopy based method by combining the entropy ratio and the root-square-mean error of low bands was proposed. When the discrete wavelet transform was applied in the signal decomposition, the wavelet entropy at decomposed levels was calculated, then the entropy ratio and the root-square-mean error of low bands were used as the deciding threshold to choose the optimal decomposi-tion level. The signal-to-noise ratio, the root-square-mean error, the relative error of peak value and the error of the peak position were used to evaluate the performance of the method. Simulation and experimental results show that the proposed method can adap-tively determine the optimal decomposition level, remove the ultrasonic signal noise and keep the useful information effectively.