Infrared spectral data denoising method based on stationary wavelet transform

摘要

For the sake of effectively alleviating the effect of noise in infrared spectral data, a method of infrared spectral data denoising based on stationary wavelet transform is proposed in this paper. Firstly, stationary wavelet transform is adopted to decompose the original infrared spectral data, which extracts data of multi-scale specific characteristic. Secondly, according to difference between spectral signal and noise in different scales, the improved variational method is introduced to adjust each sub-band coefficients. Finally, denoised signal was reconstructed through inverse stationary wavelet transform. Several groups of experimental results are demonstrated that the proposed method not only effectively extract noise but also decreases Mean Squared Error and preserve character of signal. It can be utilized in the actual infrared spectral data denosing and achieved perfect effectiveness.