紫外-可见吸收光谱法水质检测系统的噪声分析与处理研究

Noise Analysis and Denoising Research on the UV-Visible Absorption Spectroscopy Water Quality Detection System

直接紫外-可见光谱法检测水质参数具有检测速度快、成本低、无二次污染、可实现在线和原位测量等优点,但受限于光谱检测精度和检测设备成本。基于该方法的检测系统一直较难推广。为此,利用海洋光学的工业级光谱仪USB2000+及脉冲氙灯光源搭建了一套低成本、快速的水质光谱检测系统。进而,在对系统噪声源进行全面分析的基础上,利用小波包软阈值去噪方法对原始光谱数据进行去噪处理。为了有效地评价去噪效果,同时采用海洋光学科研级光谱仪QE65000配合稳定的氘钨灯光源搭建标准光谱检测系统,对同一样本溶液进行了紫外-可见吸收光谱采集。实验结果表明,搭建的系统能稳定、有效地采集原始光谱数据,进一步对原始数据采用基于软阈值的小波包去噪,能在保留水样紫外-可见吸收光谱特征的前提下,有效地去除原始光谱的各种噪声,提高光谱的信噪比,为采用低成本光谱检测设备进行水质参数分析奠定了基础。

Using direct ultraviolet-visible （UV-visible） spectroscopy method to detect water quality parameters has some advantages such as rapid detection speed, low cost, no secondary pollution, on-line and in-situ measurements. However, the detection system based on this method has been difficult to promote be cause of the limitation of to the accuracy of spectral detection and the cost of equipment. In this paper, we use the industrial-grade USB2000＋ spectrometer and pulsed xenon light source of Ocean Optics to build a low- cost, rapid water quality spectral detection system. Furthermore, on the basis of a comprehensive analysis of the system noise sources, we use the soft thresholding wavelet packet method to denoise the raw spectral data. In order to effectively evaluate the denoising effect, we use scientific-grade spectrometers QE65000 of Ocean Optics with stable deuterium tungsten light source to build a standard spectral detection system, and carry out UV-visible absorption spectrum acquisition on the same sample solution. Experimental results show that the proposed system can capture the raw spectral data stabl and effectively. Then using soft thresholding wavelet packet to denoise the original data, on the premise of retaining water samples UV-visible absorption spectrum characteristics, various noises of the original spectra are effectively removed and the signal-to-noise ratio, which makes a foundation for the low-cost spectral detection device for the analysis of water quality parameters.