Analysis Of Sesame Oil Quality And Lignans Content Based On Spectral Technology

Sesame is one of the ancient oil crops.Its sesame oil has high nutritional value and significant health benefits.It has a strong aroma and is loved by consumers at a relatively high price.However,the quality of sesame oil products in the market is uneven,and the conventional detection methods are very complicated and can not be detected at any time,so a fast and convenient detection method is urgently needed.Modern spectral technology can be non-destructive,portable,rapid detection,in the food and other industries widely used.Therefore,this study takes sesame oil as the research object,and uses spectral technology to carry out a series of studies from ensuring the authenticity of sesame oil to quality safety and quality improvement.The following are the research contents:（1）The qualitative identification and quantitative detection models of binary and ternary adulterated sesame oil were established based on Raman spectroscopy.In order to optimize the model,different spectral preprocessing methods are used for comparison,and the screening results show that the modeling effect is better without preprocessing,and then the orthogonal partial least squares discriminant analysis is established.OPLS-DA qualitative detection model,calculated the spectral characteristic variables through the model,tried to establish a partial least squares（PLS）quantitative model to screen the characteristic bands according to the selected preprocessing methods and characteristic variables,and selected the best calibration band as follows:The root mean square error of cross-validation（RMSECV）values from 203.53-1016.38 cm^-1 and 1017.35-2000 cm^-1 determined that the number of principal factors of the model 6 was the best.The results show that the R² and Q² of the OPLS-DA qualitative detection model are 0.99,and the replacement detection verifies that the model is not overfitting.The recognition accuracy of the prediction set is 100%,and the prediction ability and reliability of the PLS quantitative detection model are 0.99.the root mean square error of prediction（RMSEP）was 5.81%,the ratio of prediction to deviation（RPD）was 5.53,and the root mean square error of prediction was 0.06.The prediction results have high credibility and accuracy.This paper realizes the rapid and non-destructive qualitative and quantitative detection of binary and ternary adulterated sesame oil by Raman spectroscopy technology,and the accuracy is high.It is expected to be used as a means of market supervision and management of sesame oil.（2）Based on Raman spectroscopy,two PLS quantitative detection models were established for the acid value and peroxide value of sesame oil with different oxidation degrees.By comparing the influence of different spectral preprocessing methods on the models,smooth fitting pretreatment and smooth fitting combined with first derivative method were selected as the optimal spectral preprocessing methods for the analysis of acid value and peroxide value,respectively.Then,the competitive adapative reweighted sampling（CARS）method is used to select the feature variables of the two models,and the feature variables are converted into feature bands.The quantitative models for acid valence and peroxide value were obtained with 11 and 18 characteristic bands,respectively,and the optimal number of principal factors for the two models was determined to be 8 and 5,respectively,using RMSECV values.Finally,the R² of the prediction set of the acid value PLS quantitative model was 0.96,RMSEP was 0.11 mg/g,and RPD value was 3.02.The R² of the peroxide value PLS quantitative model was 0.99,RMSEP was 0.06 g/100g,and RPD was 3.39,which proved that the prediction ability and stability of the two models were good.The rapid quantitative detection of acid value and peroxide value in sesame oil based on Raman spectroscopy technology is realized,and it is expected to be used to control the quality and safety of sesame oil.（3）Three PLS models for rapid detection of sesamin,sesamolin and sesame lignans in sesame oil was established based on near-infrared spectroscopy.For the optimization of these three models,different spectral preprocessing methods were compared,and the best preprocessing conditions were the second derivative,the standard normal variate transformation（SNV）,and the combination of the first derivative and SNV.According to the different characteristic peaks,the spectrum was divided into three bands,and the influence of spectral band selection on the performance of the model was explored respectively.The optimal bands were 1300-1700 nm,1300-1700 nm and 110-1700 nm,and the optimal principal factor numbers were 12,26 and 10 by RMSECV value.The R² of the final PLS quantitative model of sesamin,sesamolin and sesame lignans were 0.99,0.97 and 0.94,RMSEP were 2.69μg/mL,3.73μg/mL and 7.96μg/mL,and RPD values were 7.71,6.30 and 4.10,respectively.The quantitative detection of sesamin,sesamolin and sesame lignans in sesame oil based on near infrared spectroscopy technology has been realized,and the detection speed is fast and the prediction accuracy is high.The above research results have important scientific significance for the rapid non-destructive testing of sesame oil quality safety and quality improvement,and provide important guidance and application value for the development of real-time testing equipment...