模式识别在波形分析中的应用

Application of Pattern Recognition in Waveform Analysis

根据波形自动分析的需要 ,将典型波形进行傅立叶变换 ,所得数据形成波形模式 ,继而形成典型波形空间。然后利用空间模式识别方法 ,把测试波形模式与典型波形空间中的模式进行比较 ,按离差度给出测试波形与典型波形的匹配程度 ,从而确定测试波形类型。文中还讨论了加权法模式识别在测试波形出现噪声时的应用。最后用实际应用中的波形数据进行了分析验证 ,结果表明空间模式识别方法在波形分析中具有良好性能。

In accordance with the requirement of waveform auto analysis,the typical waveforms have had fourier transformations from which data are obtainted to form waveform patterns, and further to form typical waveform pattern space. And then, the space pattern recognition method is used to compare the test waveform patterns with the models in the typical waveform space so that the matched degrees are given to the tested waveforms and typical waveforms in terms of disparity degrees whereby determing the types of test waveforms. Also,this paper discusses the application of weighted method of pattern recognition in the test waveforms with noise occured. Finally,the waveform data from practical application are used to carry out analytical tests,indicating that the space pattern recognition method has the better performances in waveform analysis.