基于最优小波包基的模极大值法信号消噪研究

给出了对信号进行基于熵准则的最优小波包基分解的基本原理,在此基础上提出了使用模极大值法进行信号消噪的基本原理.通过计算Lip指数确定信号中的奇异点,然后根据噪声和有用信号的模极大值随着小波分解尺度的增大而呈现出的截然不同的特性剔除噪声,从而实现消噪.通过对一含有噪声的信号进行消噪验证了该方法的优越性.

Study on Signal Denoiseing of Modulus Maximum based on optimail wavelet package decomposition

The basic principle of optimal wavelet package decomposition based on entropy criterion was introduced. By means of wavelet package decomposition, a signal denoise method which adopts modulus maximum was presented. The singularity points of signal are searched by calculating Lip values. With the increase of wavelet decomposition scale, the modulus maximum values of noise changes in the reverse direction to that of useful signal and this property is used to decompose noise. Simulation result demonstrates the denoise effect of the said method is better than other denoise methods.