基于极大似然估计的奇异点检测  

Singularity Detection Based on Maximum Likelihood Estimation

作　　者：刘密歌[1] 李小斌[2] 尹月琴 

机构地区：[1]西安文理学院,陕西西安710065 [2]西安电子科技大学,陕西西安710065 [3]延安市供电局,陕西延安716000

出　　处：《现代电子技术》2011年第11期82-84,共3页Modern Electronics Technique

基　　金：中央高校基本科研业务费专项资金资助项目(K50510700008)

摘　　要：信号的奇异点往往包含信号比较重要的信息,对其进行检测和定位在许多实际问题中都有重要的意义。根据脉冲型奇异点的特点,提出了一个基于极大似然估计的脉冲奇异点检测方法。该方法利用Bernoulli概型来建模,用于对脉冲奇异点的检测,将脉冲奇异点的检测问题转化为一个参数估计问题,并利用极大似然估计方法给出了参数的估计。通过对大量的仿真信号及真实信号脉冲奇异点的检测证明,该检测方法计算简单,行之有效。The singular point often contains more important information.Its detection and localization have important practical significance.According to the characteristics of pulse-type singular point,a novel method for detecting the pusle singularity is presented by using the maximum likelihood estimation（MLE）.The modelling of the singularity detection process is conducted with the independent Bernoulli probability model.The detection problem of singular points is converted to the problem of parameter estimation.The estimation of parameters is offered by using the maximum likelihood estimation method.The experimental results show that the method can indeed identify the pulse singularity from signals successfully and effectively.

关 键 词：脉冲型奇异点 Bernoulli概型 极大似然估计 奇异点检测 

分 类 号：TN911.23[电子电信—通信与信息系统]