基于匹配追踪的信号去噪方法性能影响因素分析
Analysis of Signal Denoisng Performance Influencing factors Based on Matching Pursuit
蒋柏峰1
作者信息
- 1. 中国电子科学研究院,北京 100041
- 折叠
摘要
基于匹配追踪(Matching Pursuit)的信号去噪方法的基本思想是利用信号、噪声与原子库相关性大小的差异.在实际应用中,相关性大小要用内积来衡量.当信噪比较低、信号长度有限时,内积不再准确反映相关性大小,去噪方法的性能严重下降甚至失效.该文首先介绍了此类方法的基本思想和处理流程,在此基础上深入分析了信噪比、信号长度对相干比阈值及去噪性能的影响,通过仿真验证了分析结果的正确性,并从中得到了一些有益的结论.
Abstract
The basic idea of signal denoising method based on matching pursuit(MP)is utilizing the correlation variance among signal,noise and dictionary.In practical applications,the correlation level is measured with inner product,which may lost its accuracy when SNR is low and signal length is limited.This paper firstly introduces the basic idea and procedure of such method,then analysis the effect of SNR and signal length to correlation ratio threshold and denoisng performance.Simulation results show the effectiveness of the analysis results,and provide some significant conclusions.
关键词
信号去噪/匹配追踪/相干比/内积Key words
signal denoising/matching pursuit/correlation ratio/inner product引用本文复制引用
出版年
2024
NETL
NSTL
万方数据