|
|
Research on feature extraction methods based on Singular Value Decomposition |
Duan Xiangyang;Wang Yongsheng;Su Yongsheng |
(College of Naval Architect and Marine Power,Naval University of Engineering,Wuhan 430033,China) |
|
|
Abstract Abstract: Singular values reflect the energy distribution of wanted signal and noise, the characteristic signal which implicates in noise can be extracted by singular value decomposition (SVD). The feature extraction method in strong background noise based on SVD was presented. The research shows that the distribution of singular values tends to a line with the lowerrng of Signal Noise Ratio (SNR), so the characteristic signal is difficult to separate and extract. The distribution range of the singular values which reflect the energy of noise is extended by increasing the singular value decomposition order , so the energy of noise is dispersed, and the singular values which reflect the energy of wanted signal is well-marked, and so it is favourable to the extraction of the characteristic signal. The feasibility of the method are validated by simulation testing and fault analysis example.
|
Received: 03 November 2008
|
|
|
|
|
|
|