%0 Journal Article
%T 强噪声背景信号的Perona-Malik扩散滤波算法<br>Perona-Malik Diffusion Filtering Algorithm for Mechanical Vibration Signals in Strong Noise Background
%A 毋文峰
%A 陈小虎
%J 机械科学与技术
%D 2018
%X 为了提取强噪声背景下机械振动信号的微弱故障特征，提出利用Perona-Malik非线性各向异性扩散滤波模型来实现强噪声背景信号降噪的方法。首先阐述了偏微分方程和Perona-Malik扩散滤波模型在图像降噪中的应用；其次分析了小波变换等传统信号降噪方法的不足；最后基于图像降噪和信号降噪原理的相似性，利用Perona-Malik扩散滤波模型来实现机械振动信号的降噪，将其用于轴承振动仿真信号和实测信号。实验表明，与小波阈值去噪算法等传统信号降噪方法相比，Perona-Malik扩散滤波模型更适用于强噪声背景信号降噪，同时兼顾了信号去噪和保留信号细节特征的双重要求。<br>The signal denoising preprocessing is very important for extracting weak fault features from mechanical vibration signals in strong noise background. However for wavelet transform (WT) and other traditional signal processing algorithms, the processing of signal denoising is a troublesome thing. The balance between signal denoising and feature preserving is a couple of contradictions. Thus we mean to bring in the Perona-Malik nonlinear anisotropy diffusion filtering model to process signal preprocessing in strong noise. Firstly, the partial differential equation (PDE) theory is introduced, and the Perona-Malik model, as one of important nonlinear anisotropy diffusion filtering algorithms, is brought in. Secondly, from the applications in image denoising, it can be analyzed and inferred that the Perona-Malik model is a perfect solution for wavelet transform and other traditional signal denoiseing algorithms. Lastly, by comparison with the wavelet threshold denoising algorithms in the bearing vibration signals, it can be indicated that the Perona-Malik model is very appropriate for mechanical vibration signals in strong noise. Above all, the Perona-Malik filter can not only realize signal denoising but also preserve signal features with better denoising performance and without signal distortions
%K 小波变换
%K 偏微分方程
%K Perona-Malik模型
%K 信号处理
%K 信号降噪&lt
%K br&gt
%K wavelet transform
%K partial differential equation
%K Perona-Malik model
%K signal processing
%K signal denoising
%U http://journals.nwpu.edu.cn/jxkxyjs/CN/abstract/abstract7088.shtml