Research on Threshold Improved Denoising Algorithm Based on Lifting Wavelet Transform in UV-Vis Spectrum
Zhou Feng-bo,Li Chang-geng,Zhu Hong-qiu
DOI: https://doi.org/10.3964/j.issn.1000-0593(2018)02-0506-05
2018-01-01
Spectroscopy and spectral analysis
Abstract:In the quantitative analysis of UV-Vis spectroscopy, the random noise which affects the accuracy of the spectral quantitative analysis results seriously, is caused by spectrophotometer internal optical systems, light sources, detectors, electronic components, circuit design and external environmental interference and other factors. In order to improve the accuracy of UV-Vis spectral analysis, the spectral data need to be denoised firstly. Because wavelet analysis has the characteristics of multiresolution, low entropy and decorrelation, the denoising algorithm based on wavelet transform is superior to the traditional denoising algorithm. Currently there are threshold denoising method, coefficient correlation denoising method and modulus maxima denoising method in wavelet transform domain. In these method, the threshold denoising method proposed by Donoho is the most commonly used in engineering application. According to Donoho threshold denoising principle, this paper proposes a threshold improved algorithm based on lifting wavelet transform. On the one hand, the lifting wavelet transform is the second generation wavelet transform, inherits the multi-resolution characteristic of the wavelet transform, and does not need the Fourier transform, which has the characteristics of small computation, fast speed and simple realization. On the other hand, a new threshold function is proposed to overcome the discontinuous shortcoming in the hard threshold method and reduce the constant deviation in the soft threshold method. At the same time, the threshold estimation is adjusted to facilitate the retention of the signal wavelet coefficients and the elimination of the noise wavelet coefficients. The UV-Vis spectrum of three groups poly-metal ions were used to test the performance of the proposed denoising method. In the experiment, random noise was first added to the spectrum, and then removed by the proposed denoising method. The signal to noise ratio(SNR) and the root mean square error(RMSE) were used to evaluate the performance of the proposed denoising method. The experimental results showed that the proposed method was superior to the soft threshold method and the hard threshold method in improving SNR and decreasing RMSE, which can effectively eliminate the spectral noise and keep some important detail features in the spectral signal. So, this proposed method is more suitable for the denoising pretreatment before the UV-Vis spectral data modeling, and will have a good application prospect in UV-Vis spectroscopic analysis.