信号自适应去噪方法的仿真研究  

Study and Simulation on Signal Adaptive De-noising Method

作　　者：陈华丽[1] 刘康[2] 程耕国[1] 

机构地区：[1]武汉科技大学冶金自动化与检测技术教育部工程研究中心,湖北武汉430081 [2]电子科技大学电子工程学院,四川成都611731

出　　处：《计算机仿真》2011年第1期344-347,387,共5页Computer Simulation

基　　金：国家自然科学基金(60672064);武汉科技大学科学研究发展基金(2006XZ3)

摘　　要：研究信号问题,实际中信号都带有噪声。对不同的信号寻找最佳的去噪方法一直是信号处理和检测的主要问题,传统的信号去噪方法存在基函数单一,或者基函数难以选择的问题,使去噪效果不理想。提出一种新的基于Hilbert-Huang变换的自适应的信号去噪方法,解决了传统去噪方法存在的问题,提高了信号去噪的效果。方法是一种新的分析非线性非平稳信号的时频方法,包括经验模态分解(EMD)和Hilbert变换两部分,从信号本身的尺度特征出发对信号进行EMD分解,得到一组固有模态函数,具有良好的局部自适应性。进行仿真证明,方法的基函数具有自适应性,能很好的匹配信号的特征,既能分析平稳信号又能分析非平稳信号,尤其是对短时的非平稳信号进行去噪是非常有效的。Finding a better de-noising method for different signal has been one of the major issues in the signal processing and detection field. The traditional methods have some disadvantages that make the de-noising effect not i- deal such as basis function is too singular, or the choice of the basis function is too difficult etc. The new adaptive signal de-noising method is advanced based on Hilbert-Huang Transform which overcomes the disadvantages of tradi- tional de-noising methods and improves the signal de-noising effect. The method is a new time-frequency method analyzing non-stationary and nonlinear signals, which includes empirical mode decomposition （ EMD ） and Hilbert transform. According to the scale characteristics, the signal was made EMD into a series of IMFs in the method with good local adaptive. MATLAB simulations prove that the basis functions in the new method are adaptive and can match the signal characteristics much better, which can not only analyze stationary signal, but also analyze non-sta- tionary signal analysis, and is a very effective method especially for short-term non-stationary signals de-noising.

关 键 词：希尔伯特-黄变换 小波分析 经验模态分解 固有模态函数 

分 类 号：TN911.23[电子电信—通信与信息系统]