| The quality and safety of water are related to national economy and people’s livelihood.From time to time,incident of organic matter pollution in water occurs and the people’s health and social harmony are seriously affected.It is of great significance to explore an effective method for the detection and classification of organic pollutants in water.Ultraviolet-visible(UV)spectrum and three-dimensional(3D)fluorescence spectrum have attracted much attention in the qualitative detection,identification and alarm of water quality organic pollutants due to their high sensitivity,high selectivity,no reagents and fast detection speed.This thesis focuses on the actual needs of scientific research projects and urban water supply supervision departments,and focuses on issues such as the impact of environmental factors on spectral detection methods,improving the detection performance of organic pollutants beyond standards,and improving the classification of characteristic organic pollutants.This thesis has studied the techniques of spectral analysis,feature extraction and modeling,UV-fluorescence information fusion,over-limit detection,identification and classification in the detection of organic pollutants in raw water by UV spectrum and 3D fluorescence spectrometry.An organic pollutant detection and analysis method based on the fusion of characteristic information of UV spectrum and 3D fluorescence spectrum is explored and verified by experiments.The main work and innovations of this thesis are summarized as follows:(1)The detection principle of water quality detection based on UV spectrum and 3D fluorescence spectrum is studied,and the characteristics,differences and information complementarity of UV spectrum and 3D fluorescence spectrum are analyzed.The feasibility of fusion modeling using the characteristic information of these two spectra is studied.Experiments and modeling analysis are carried out by taking organic such as phenol,aniline,and salicylic acid as examples,and the feasibility verification is performed.(2)The influence mechanism of environmental factors on the detection results of UV spectrum and 3D fluorescence spectrum is studied,and experimental analysis and verification are carried out.The effects of turbidity and pH changes on UV spectrum and 3D fluorescence spectrum are studied.The experimental results show that the changes in turbidity have a great impact on the UV spectrum and the 3D fluorescence spectrum,and the change rules at different bands are difficult to accurately quantify.Sample filtering and data SG smooth filtering pretreatment are used to reduce the turbidity on the UV spectrum and 3D fluorescence spectrum.The effect of pH change on the spectrum is mainly the effect of strong acidity and alkalinity on the spectrum.The threshold method can be used to reduce the effect of pH change.Through water sample processing and data processing,it provides guidance for subsequent research on the detection and classification of organic pollutants based on UV-Vis and 3D fluorescence spectra.(3)The detection method of whether the water organic matter exceeds the limit based on the fusion of UV-fluorescence spectrum information is studied.Aiming at the problems of high detection limit of organic pollutants in the UV spectrum single-spectrum model and the high false-positive rate of the 3D fluorescence spectrum single-spectrum model due to the normal fluctuation of raw water quality,a method for detecting organic matter in water based on the fusion of UV-fluorescence spectrum is proposed.For UV spectrum,wavelet transform and PCA(Principal Component Analysis)are used to extract features.After preprocessing the 3D fluorescence spectrum,2D wavelet transform and energy coding are used to extract features.At the feature level,the UV principal component is integrated with the energy encoding features of the fluorescence spectrum to establish a discriminant model.Together with the single spectrum model,a discriminant model set is formed,and then a vote is used to determine whether the organic molecular weight concentration exceeds the upper limit.The experimental results show that the UV-fluorescence spectrum fusion model can effectively reduce the detection limit and reduce false negatives than the UV-Vis single-spectrum model;compared with the 3D fluorescence single-spectrum model,the UV-fluorescence spectrum fusion model can effectively reduce the false alarm rate.(4)A method for classifying and identifying water organisms based on UV-fluorescence spectral information fusion is studied.The similar response of some organic compounds in UV spectrum or 3D fluorescence spectrum results in poor classification performance.Aiming at this problem,a method for identifying organic pollutants based on the fusion of UV spectrum and 3D fluorescence spectrum information is proposed.This method can effectively use the difference information of substances in different spectra to achieve substance discrimination.For the UV spectrum,the competitive adaptive re-weighting method is used to extract the specific band of the UV spectrum.For the 3D fluorescence spectrum,the combination of the non-downsampling Contourlet transform and the gray level co-occurrence matrix are used to extract the 3D fluorescence spectrum features.Finally,multi-class support vector machines are used to achieve classification.The experimental results show that the spectral fusion model can effectively distinguish substances with similar spectra in a single spectrum.Compared with the single-spectrum model,the classification and recognition of substances is better.In general,this thesis focuses on the detection and classification of organic pollutants in raw water of urban water supply.The method of comprehensively using UV spectrum and 3D fluorescence spectrum information to detect and identify organic pollutants is studied.Theoretical analysis and experimental results show that the spectral feature information fusion modeling method uses the complementary information of the two spectra,which can improve the performance of detection and classification of over-limits,and has reference significance for urban water supply monitoring and early warning. |
PDF Full Text Request