基于小波去噪的阈值函数改进方法分析

摘要

首先介绍了两种国内外比较有代表性的小波阈值改进方法,并在传统软、硬阈值函数的基础上提出了一种新的改进阈值函数,通过对小波系数的阈值处理,使得去噪后的信号在过分平滑与边缘振荡现象之间达到合理的平衡.对一段压气机台架实验数据的仿真分析表明:该方法在去噪的同时减少了有用信息的损失,在信噪比增益和最小均方差误差方面的性能指标要优于传统软、硬阈值法,同时在3种改进阈值函数方法中具有最佳的去噪性能.%Two typical methods for modifying wavelet threshold functions were introduced,and basing on tra-ditional soft and hard threshold functions,a new modified threshold function was proposed.Through the threshold processing of wavelet coefficients,the signals de-noised reached the balance between excessive smoothness and edge oscillations.Simulating experiment data from a compressor bench shows that this method can reduce the loss of useful information while de-noising the signals,and the performance index is superior to traditional soft and hard in SNR gain and in the minimum mean square error,and it has the best de-noising performance of three modified threshold methods.