Chinese Journal of Quantum Electronics, Volume. 24, Issue 3, 294(2007)
Noise-filtering of surface plasmon resonance spectra with wavelet analysis
Based on optical fiber surface plasma resonance sensor,the application of wavelet denoising with threshold for the SPR reflected spectrum is studied. At first,Gaussian white noise is added to the theoretical spectrum,and five kinds of mother wavelets,which are Daubechies9,Symlet7,Symlet11,Symlet14 and Symlet15 are selected to reduce the noise. Comparing the resonant reflectivity and wavelength of SPR spectrum after denoising with that of the theoretical value,Symlet11 wavelets obtained the best performance. Then,five groups of the experimental spectrum data were acquired from optical fiber SPR sensing system at different time,and their noise are reduced by Symlet11 wavelets. The standard deviations of experimental resonant wavelength and reflectivity after denoising are 0.4039 nm and 0.2231%. All these work proves that wavelet denoising with threshold can well correct the migration of resonant wavelength caused by noise,and it will bring convenience to the practical application of optical fiber SPR sensor.
Get Citation
Copy Citation Text
D U Yan, LIANG Da-kai. Noise-filtering of surface plasmon resonance spectra with wavelet analysis[J]. Chinese Journal of Quantum Electronics, 2007, 24(3): 294
Category:
Received: Apr. 19, 2006
Accepted: --
Published Online: Jun. 7, 2010
The Author Email: Yan D U (duyan_nuaa@163.com)
CSTR:32186.14.