Study On Detection Of Food Pigment Based On Support Vector Machine And Fluorescence Spectroscopy

The improvement of living standards is constantly updating people's requirements for food.Food with bright color and better taste is often more popular in the complex food group.So many bad businesses appear to use food pigments in violation of the rules.Food pigments,also known as colorants,include synthetic and natural pigments.Compared with natural pigment,synthetic pigment has more brilliant color,stronger colouring ability and lower price.It is widely used in food,industry,medicine and cosmetics and so on.However,synthetic food pigments are generally toxic and resulting in malformations.Due to excessive intake,the cancer cases emerged in endlessly.The state has strict regulations on the use of various food pigments.Therefore,selecting an efficient detection method is significant to solve the food safety problem.In this paper,three-dimensional fluorescence spectroscopy combined with optimized support vector machine(SVM)algorithm was used to detect the substance and concentration of synthetic food pigment,and the qualitative and quantitative analysis of artificial food pigment was realized.The main research contents are as follows:(1)Combined with the material structure of artificial synthetic pigments and the principle of photoluminescence,the feasibility of using three-dimensional fluorescence spectroscopy to analyze artificially synthesized food coloring materials was analyzed.Fluorescence spectra of different concentrations of carmine,amaranth and tartrazine samples were obtained by FS920 steady-state fluorescence spectrometer.And their fluorescence peaks and fluorescence intensity were compared.The fluorescence spectra of mixed solution of three kinds of food pigments were measured,and the fluorescence characteristic peaks and fluorescence intensities of mixed solutions of two kinds of pigments and three kinds of pigments were studied and compared with the fluorescence spectra of monochromatic food pigment solutions.And then analyze their effects on each other.Because of the complex interaction among the three pigments,the qualitative and quantitative analysis of pigment can not be carried out only by using three-dimensional fluorescence spectroscopy.It is necessary to establish anaccurate detection model by combining chemometrics and other algorithms to classify the pigment species and predict the concentration.(2)The improved particle swarm optimization algorithm is used to optimize the parameter optimization process of support vector machine,and the detection model of optimized support vector machine is established.The detection model is used to distinguish the three-dimensional fluorescence data of mixed solution of three kinds of food pigments.The results show that the optimal support vector machine detection model can accurately classify the mixed pigment solution,and the effect is obviously better than the detection model before the optimization of the algorithm.(3)Simulated the complex environment of multi-pigment combination in food,the content of carmine in mixed solution of three pigments was detected by optimized support vector machine algorithm.The high recovery rate and low root mean square error show that the detection model can quantitatively analyze the pigment in mixed solution under the presence of interference pigment,and the results show obvious advantages compared with other detection models.(4)The established model was used to detect the content of carmine in carbonated drinks,fruity drinks and cocktails,and the standard addition method was used to verify the results.The optimized support vector machine algorithm combined with three-dimensional fluorescence spectroscopy can be used to identify and quantitatively detect food pigments in mixed system.