Research On Data Processing Method Of Water Quality Detection Based On UV-Vis Spectroscopy

The rapid development of China's economy is always accompanied by an unoptimistic situation of water quality.Lucid waters and lush mountains are invaluable assets.We must treat the ecological environment as we do for life.Therefore,research on water quality detection technology cannot be delayed.The traditional water quality detection technology has the disadvantages of complicated operation,long measuring period,need of chemical reagents and secondary pollution.However,the direct ultraviolet visible spectrometry water quality detection technology has the advantages of real-time,in-situ,on-line measurement,no chemical reagents,no secondary pollution,etc.It has been widely used in the field of online water quality detection.However,there are still some deficiencies in data processing of water quality detection technology by direct UV-Vis spectrometry.For example,the water quality detection system is affected by the measurement instrument itself and the external environment,resulting in spectral data containing a lot of noise,and a single model is not suitable for spectral analysis of various waters,which seriously affects the accuracy and analysis accuracy of water quality spectral data modeling.Based on this,the thesis,with the joint funding of the Chongqing University's postdoctoral scientific research project special support(Xm2016117)and the Chongqing Municipal Education Commission's science and technology research project(KJ1709201),focuses on the key technical issues in the spectral data processing of the water quality online detection technology,and deeply researches on data processing method of water quality detection system based on UV-Vis spectroscopy.The main research work of the dissertation is as follows:1)UV-Vis spectrometry water quality detection system design.Using a deuterium tungsten halogen lamp combination light source DH2000 and a scientific-grade spectrometer QE series,a UV-Vis spectrometry-grade water quality detection system was set up to collect domestic sewage,stream water,and pure water as water quality samples,and the UV-Vis spectral data of the samples were obtained.2)Research on denoising of UV-Vis spectrometry water quality detection data by compressed sensing algorithm based on wavelet transform.Aiming at the serious problem that water quality detection spectrum detection system is affected by light source,opticalpath,photoelectric conversion device and complex environment,this paper proposes a compressive sensing denoising algorithm based on wavelet transform.This paper focused on the comparison of denoising effects of compressive sensing algorithm with different compression ratios and different observation matrices,and compared compressive sensing denoising method with the traditional wavelet threshold denoising method.The experimental results show that the compressed sensing denoising algorithm based on wavelet transform can effectively remove the system noise.3)Research on dimension reduction method of water quality detection spectral data by UV-Vis spectroscopy with principal component analysis(PCA).For the problem that the spectral data collected by the water quality inspection system of direct spectroscopy has large information redundancy and complicated analysis process,this paper proposes a principal component analysis method.The principal component analysis method not only eliminated the correlation among the information indexes,realized the dimension reduction of the spectral data and the feature information extraction,but also simplified the data structure and shortened the data analysis time.4)Research on discriminant classification of water quality detection spectral data by UV-Vis Spectroscopy with Fisher discriminant method.In response to the problem of poor adaptability and low resolution of traditional single water quality spectral data model,this paper proposes to establish a discriminant model using PCA combined with Fisher discriminant method.The discriminant model was used to classify a chemical plant and a stream water sample,and compared with the traditional Fisher discriminant model.The experimental results show that PCA combined with Fisher discriminant model can effectively eliminate the influence of information redundancy,with good classification effect,low misjudgement rate and short computation time.The compressive sensing denoising algorithm proposed in this paper provides a new idea for the denoising of spectral data in real-time online water quality detection system,and has important guiding significance for instrument noise reduction.The method of principal component analysis combined with Fisher discriminant provides an efficient method for practical engineering of direct spectroscopy water quality type discrimination.