首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于D-J阈值的加性白噪声方差估计改进方法研究
引用本文:韩骏,何敬锁,刘菁.基于D-J阈值的加性白噪声方差估计改进方法研究[J].首都师范大学学报(自然科学版),2007,28(2):36-39,43.
作者姓名:韩骏  何敬锁  刘菁
作者单位:1. 首都师范大学教育技术系,北京,100037
2. 都师范大学物理系,北京,100037
摘    要:本文提出了一种基于D-J阈值的加性白噪声方差估计方法.与中值方差估计方法相比,采用本文提出的方法估计的噪声方差从统计意义上看更准确,并具有更小的离散性.本文通过模拟的逐段光滑信号来评价该方法的有效性,并给出了与中值方差估计方法的比较结果.结果显示特别是在信号采样序列的长度有限时,采用本文方法可以获得比中值估计方法更准确、更可靠的白噪声方差估计.

关 键 词:方差估计  加性白噪声  小波变换  D-J阈值
修稿时间:2006-07-04

D-J Threshold-Based Estimation of White Noise Variance
Han Jun,He Jingshuo,Liu Jing.D-J Threshold-Based Estimation of White Noise Variance[J].Journal of Capital Normal University(Natural Science Edition),2007,28(2):36-39,43.
Authors:Han Jun  He Jingshuo  Liu Jing
Institution:1.Department of Educational Technology, Capital Normal University, Beijing 100037; 2 Physics Department, Capital Normal University, Beijing 100037
Abstract:In this paper,a Donoho and Johnstone(D-J) threshold-based approach for the estimation of variance of additive white noise corrupting a signal is presented.Compared with the median variance estimator,the variance estimated using this approach is more accurate and of smaller divergence in a statistical sense.In particular,when the sample sequence of a signal has a limited number of samples,the approach provides more reliable estimate of noise variance than themedian variance estimator.Evaluation of the approach is implemented using a simulated piecewise smooth signal.Comparisons with the median variance estimator are also given.
Keywords:Variance estimation  additive white noise  wavelet transform  D-Jthreshold  
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号