Study On Detection Method Of Pesticides Residues Based On Fluorescence Spectroscopy Mechanism

Pesticide residues in water and agricultural products will damage the ecological environment and endanger human health.Nowadays,the problem of pesticide residues has become a global focus issue.Fluorescence spectroscopy is regarded as a rapid,high sensitive and precision,simple operation and low cost method which could be used to detect pesticides remaining in water and agricultural products.It is of great significance for maintaining ecological balance,ensuring the safety of agricultural products and human health.This paper aims is to solve the problems of detecting multiple trace pesticides mixed in the water environment and agricultural products.It combines fluorescence spectroscopy technology with missing data recovery and chemometrics to eliminate the first-order Rayleigh scattering in the spectrum and the mixture of pesticides in water environment were accurately identified and predicted;different artificial neural networks were carried out to do quantitative analysis in two common typical agricultural products,further discussed the experimental results.The main work of this paper is as follows:1.Based on the density theory in quantum chemistry,the ground state and excited state geometric configurations were analyzed,the bond length,bond angle and dihedral angle of Carbaryl,Carbendazim,Tsumacide and Triazophos pesticide were discussed,the theoretical values of the fluorescence wavelength of each pesticide molecules were calculated,which provided an explanation for the fluorescence phenomenon and verified the feasibility of fluorescence spectroscopy for the detection of pesticides.2.In view of the problem that the first-order Rayleigh scattering is difficult to remove in the fluorescence spectra,and the phenomenon that the Triazophos,Carbendazim and Carbaryl cause serious spectral overlap due to the similar chemical structure,missing data recovery was proposed and used to remove first-order Rayleigh scattering,which is expected to become a new data processing method.Then combined with the second-order calibration to do accurate qualitative and quantitative analysis,and calculated the parameters of correlation coefficient,root mean squared error of prediction,average recovery rate,limit of detection and quantitation.This method provides a new tool to remove the first-order Raleigh scattering from the spectrum and green,non-destructive,fast detection method for pesticide residues in water environment.3.This part is to apply the echo state network to the detection of various trace pesticides in flower agricultural products.Take honeysuckle as an analysis sample,considering that many factors will cause the fluorescence intensity and concentration in the spectral data to deviate from the linear relationship,the echo state network(ESN)was introduced to establish calculation and prediction model from the perspective of information processing,and performed quantitative analysis of pesticide residues in honeysuckle.Established echo state network,BP neural network and cyclic neural network model respectively,evaluated the error parameters of root mean square error,average absolute error and average absolute percentage error.Recovery rate and average recovery rate were also calculated.This study verifies the feasibility and effectiveness of the neural network for detecting pesticide residues in honeysuckle,good prediction results prove the advantages of the ESN,and provide a theoretical basis for related detection work.4.This part mainly studied the mechanism of BP neural network,and proposed mind evolutionary algorithm to optimize the deficiencies of BP neural network,so as to establish a new optimization algorithm and applied it to detect carbamate pesticides on fruit and vegetable agricultural products.Eight groups of parameters were compared and analyzed from the perspectives of overall prediction results and quantitative analysis respectively on tomato fruits,which clarified that the MEA-BPNN to process fluorescence spectrum data with good robustness.The influence of four cleaning methods on the degradation of pesticides was discussed to determine the best cleaning method to remove pesticides on the surface of tomatoes.This method provides a reference and guidance for the detection of carbamate pesticide residues on tomato surface in daily life.