在紫外可见光谱定量分析中, 由于分光光度计内部的光学系统、 光源、 检测器、 电子元器件, 电路设计以及外部环境干扰等因素产生的随机噪声, 严重影响光谱定量分析结果的准确性, 为提高紫外可见光谱分析精度, 需要对光谱数据进行去噪预处理。 由于小波分析具有多分辨率, 低熵性、 去相关性等特点, 基于小波分析的去噪算法优于传统的去噪算法, 目前基于小波去噪的方法主要有模极大值去噪算法, 系数相关去噪算法, 阈值去噪算法, 工程实际应用以Donoho的阈值去噪法最为常用。 根据Donoho阈值消噪原理, 提出一种基于提升小波变换的阈值改进算法, 一方面使用提升小波变换, 提升小波变换是第二代小波变换, 继承了小波的多分辨率特性, 并且不需要进行傅里叶变换, 从而具有算法简单, 速度快, 实现简单的优点; 另一方面提出了一种新的阈值函数, 克服了硬阈值函数在阈值处不连续以及软阈值函数存在恒定偏差的问题, 同时对阈值估计进行了调整, 有利于信号小波系数的保留和噪声小波系数的剔除。 对三组多金属离子混合溶液的实测紫外可见光谱信号, 添加随机噪声后使用该方法进行去噪处理, 并使用信噪比(SNR)和均方根误差(RMSE)进行去噪性能评价。 试验结果表明, 提出的算法优于Donoho的软硬阈值去噪算法, 能够有效提高光谱信噪比和降低均方根误差, 从而更好地消除光谱信号中的噪声和保留光谱信号中一些重要的细节特征, 比较适合用于紫外可见光谱数据建模之前的去噪预处理, 在紫外可见光谱信号分析中具有较好的应用前景。

In the quantitative analysis of UV-Vis spectroscopy, the random noisewhich affects the accuracy of the spectral quantitative analysis resultsseriously, 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 keepsome 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.