The de-noising method of EMD threshold based on correlation

Author

Xiaofeng, Li ; Mingjie, Liu

Author_Institution

Beijing Inst. of Technol., Beijing, China

fYear

2010

fDate

24-28 Oct. 2010

Firstpage

2613

Lastpage

2616

Abstract

This paper presents the de-noising method of EMD threshold based on correlation, using features of time scale filtering of EMD. Firstly, the sequence of noisy samples are disassembled by EMD to calculate the order of the intrinsic mode function (IMF) of them and calculate the correlation coefficient of IMF Portfolio with different order and Sampling sequence. Secondly, correlation threshold is calculated using signal to noise ratio(SNR) of sampling sequence. Signal-led IMF and noise-led IMF being distinguished using this threshold, we can define these combination of Signal-led IMF as result of de-nosing sampling sequence. The method is an adaptive approach without artificial setting parameters. In the last part of this paper, the method is used to passive acoustic localization research and achieves good results.

Keywords

acoustic signal processing; filtering theory; signal denoising; signal sampling; EMD threshold; IMF; IMF Portfolio; SNR; correlation coefficient; correlation threshold; denoising method; intrinsic mode function; noise-led IMF; noisy sample sequence; passive acoustic localization research; sampling sequence; signal to noise ratio; signal-led IMF; time scale filtering; Artificial neural networks; Correlation; Filtering; Noise reduction; Signal to noise ratio; Wavelet transforms; Empirical mode decomposition; acoustic passive localization; correlation coefficient; signal de-nosing; signal to noise ration;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing (ICSP), 2010 IEEE 10th International Conference on

Conference_Location

Beijing

Print_ISBN

978-1-4244-5897-4

Type

conf

DOI

10.1109/ICOSP.2010.5656128

Filename

5656128