信号预处理中错点剔除方法的研究

Research on the Method to Reject Abnormal Data in Signal Preprocessing

对工程信号进行错点剔除,是信号预处理中的一项重要内容.本文分析了统计学中用于错点剔除的肖维勒准则、狄克逊准则、格拉布斯准则和拉依达准则,针对上述各项准则在处理大样本信号中的局限性,引入一种新式错点剔除方法,并用拉依达准则与此方法对一段实际工程信号进行处理对比,结果表明,新式错点剔除方法能够更好地剔除错点.

To filter out the abnormal data from engineering signal is important in signal preprocessing.This paper introduced the statistical methods which are used to reject the abnormal data,including the Chauvenent＇s Criterion,the Dixon＇s Criterion,Grubbs＇ Criterion and the PAu-Ta criterion.There are limitations in using above criterions to process large sample signal,so this paper introduced a new method to reject the abnormal data,then use PAu-Ta criterion and this new method to process an actual engineering signal.The result indicates that the new method can better solve the problem.