Study On Detection Methods For Quality And Safety Of Edible Oil Based On Raman Spectroscopy Technique

The quality and safety of edible oil are closely related to public health.Conventional edible oil quality and safety detection methods are complicated to operate,expensive equipment,and high requirements for testers.Therefore,it is particularly important to explore a quick and effective method for measuring the storage quality and safety of edible oil.Raman spectroscopy has been widely used in many fields due to its advantages of fast,nondestructive and convenient operation.Therefore,this study took edible oil as the object and carried out detailed study on non-destructive detection methods for quality and safety of edible oil based on Raman spectroscopy.Specific research contents are as follows.(1)Study on detection methods of edible oil storage quality by Raman spectroscopy.Raman spectra of edible oil samples with different storage periods were collected by Raman spectrometer and the spectra were pretreated by Savitzky-Golay(SG),multiple scattering correction(MSC)and adaptive iteratively reweighted penalized least squares(air PLS).Then principal component analysis(PCA)was used to mine the features of the preprocessed spectra.Support vector machine(SVM)was introduced to establish the qualitative identification model of edible oil storage period.In model calibration,the optimization effects of grid search(GS),particle swarm optimization(PSO)and grey wolf optimization(GWO)on the parameters of the SVM were compared and analyzed.The results obtained showed that different parameter optimization methods had a significant impact on the recognition accuracy of the SVM models.When the number of principal components(PCs)was 9,the correct recognition rate of the GWO-SVM best recognition model in the prediction set reached 100%.Bootstrapping soft shrinkage(BOSS)and variable combination population analysis(VCPA)were used to select feature variables for Raman spectra after SG+MSC+air PLS preprocessing,and the SVM detection models were established based on optimized features for the acid value of edible oil in the storage process.The results obtained showed that the performance of the SVM models based on the optimized characteristic wavenumbers were significantly improved.Among them,the VCPA-SVM model had the best detection performance.The root mean square error of prediction(RMSEP)and coefficient of determination(R in prediction set were 0.153 mg/g and 0.932,respectively.The results demonstrate that it is feasible to use the Raman spectroscopy to predict the quality change of edible oil in the storage process,which provides a fast and effective analytical tool for the detection of edible oil storage quality.(2)Study on detection methods of edible oil safety based on Raman spectroscopyIn the experiment,edible oil sample containing AFB1 was prepared by squeezing peanut oil,and the obtained spectra were preprocessed by wavelet transform(WT)combined with air PLS.The PCA was introduced to mine the intrinsic characteristics of Raman spectra after pretreatment,and a qualitative identification model was established with chemometrics method to realize the qualitative identification of the AFB1 contaminated edible oil samples.The results obtained showed that the correct identification rate of the PCA-SVM model for cooking oil samples with different AFB1 pollution levels reached 100% in both the training set and the prediction set.The competitive adaptive reweighted sampling(CARS)method was used to optimize the characteristic bands of Raman spectra pretreated by WT+air PLS,and a partial least square(PLS)detection model for the AFB1 content was established based on the optimized features.The results obtained showed that the RMSEP and R of the optimal CARS-PLS model in the prediction set were 22.565 ?g/kg and 0.992,respectively.The results demonstrate that the Raman spectroscopy combined with appropriate chemometrics can be used to quickly detect the safety of edible oil with high precision.The overall results can provide a technical basis and method reference for the design and development of the portable Raman spectroscopy system for the quality and safety detection of edible oil storage,and also provide an effective and green detection tool for food regulatory authorities and grain and oil production enterprises.