| With the progress of society,the demand for high quality life is growing,and the requirement on food is the most prominent.There are more attention on intrinsic quality of fruits and vegetables in daily life,such as VC rich tomatoes.And the food safety is an enduring concern,such as the distinguishing of fermented apple vinegar and blended apple vinegar,and the identification of hogwash oil.The timely processing of above problems requires is a rapid detection technology,with convenient,no sample-pretreatment and other advantages.Near infrared spectroscopy(NIRS)is a new type of qualitative and quantitative analysis technology,it has above characteristics,widely used in agriculture,food,medicine,petroleum industry etc.In this paper,using near infrared spectroscopy analysis technical,a rapid detection of tomato nutrient composition was achieved,the linear and nonlinear model of soybean oil mixed with frying oil were studied;rapid identification method of fermented apple vinegar and blended apple vinegar was developed.The main research results of this paper are as follows:1.Through single factor analysis of variance,the effects of integral time,average number and eveness on near infrared spectral response characteristics were studied.The results show that tomatoes,apple vinegar and oil samples respectively had the minimum absorbance standard deviation when the integral time were 3 ms,5 ms,3 ms respectively;the tomatoes,apple vinegar and oil samples had the minimum relative standard deviation of absorbance,when scan times were 10 times and 15 times,10 times respectively;the tomatoes,apple vinegar and oil samples had the minimum relative standard deviation of absorbance,when eveness were 5,15 and 11 respectively.2.To realize the simultaneous detection of tomato pulp sample a variety of nutrients content by near infrared spectroscopy,principal component regression(PCR)and partial least squares(PLS)was used to establish the calibration models of tomatoes nutrient contents.And the calibration models from three chemometric methods were compared.The models established by PLS were optimal.Then near infrared spectroscopys of unknown tomatoes samples were collected under the same conditions,the correction model gave the internal components of the samples.Specific conclusions are as follows:The prediction correlation coefficient of the 4 mathematical models,tomato total acid,soluble sugar,vitamin C and protein contents were 0.951,0.970,0.974,0.987.That can meet the requirements of agricultural products detection.3.To achieve the rapid detection of frying oil quantitative and qualitative analysis by near infrared reflectance spectroscopy,the linear and nonlinear mathematical model of frying oil content was established,in the 4 kinds of correction model,the predictive ability of nonlinear regression mathematical model is higher than the linear mathematical model,in which the prediction ability of SVR statistical regression model is strongest,after the selection of interval partial least squares regression(i PLS),the validation coefficient of determination was 0.975,greater than 0.95,the root mean square error of prediction is 0.084,was smaller,meet the requirements of practical applications.At the same time,respectively,by using principal component analysis and system cluster of frying oil and soybean oil were identified,in the test 7samples of frying oil were identified.4.This paper has conducted the problems on the difficult to accurate and quick identification of blended apple vinegar and fermented apple vinegars.In the test 25 blended apple vinegar and 55 fermented apple vinegars were identified.After plectrums pretreatment of different methods and principal component analysis,the blended apple vinegar and fermented apple vinegars were distinguished,that show the near infrared spectroscopy analysis technique can be used as a fast,simple and convenient method for identification of blended apple vinegar. |
PDF Full Text Request