| With the gradual improvement of living standards of Chinese residents,people's judgment on the quality of water milled glutinous rice flour has shifted from the previous appearance sensory quality indicators such as color and odor to the internal physical and chemical indicators such as moisture,acidity and starch content.At present,in our country's quality assurance system of water milled glutinous rice flour,chemical analysis and manual judgment are still used for its internal quality detection.These detection methods have the disadvantages of large subjective factors,strong professionalism and long time-consuming.Therefore,in this study,NIR technology combined with chemometrics method was used to establish a quantitative model of internal quality index of water milled glutinous rice flour with high accuracy,good stability and strong environmental adaptability,and a rapid non-destructive testing system for internal quality of water milled glutinous rice flour with high accuracy,good stability and strong environmental adaptability,and a rapid non-destructive testing system for internal quality of water milled glutinous rice flour based on Android system was developed.The main contents and conclusions of this study are as follows:1.In order to solve the problems of professional,time-consuming and damaged samples of conventional detection methods,this chapter carried out the research on rapid non-destructive detection method of moisture and acidity of water milled glutinous rice flour.In the experiment,the NIR of the samples were collected by Texas Instruments spectrometer,and the corresponding moisture content and acidity values were detected by standard method.On this basis,the quantitative prediction model of water content and acidity was established by chemometrics method.Combined with spectral preprocessing method,the chemometric models established by ant colony algorithm(ACO),genetic algorithm(GA)and competitive adaptive reweighting sampling algorithm(CARS)combined with PLS all have good results.The results showed that the water content and acidity model established by cars-pls was the best,the correlation coefficient of water calibration Rc= 0.9973,the correlation coefficient of prediction Rp = 0.9968,the corresponding root mean square error of correction set and prediction RMSEC = 0.1733,RMSEP= 0.2126.The correlation coefficient of acidity correction Rc = 0.8256,the correlation coefficient of prediction Rp = 0.8434,the corresponding root mean square error of correction set and prediction RMSEC = 0.0701,RMSEP = 0.0703.It can be seen from the results that the model has a good prediction effect on the moisture content and acidity of water milled glutinous rice flour,indicating that NIR technology can realize the non-destructive detection of the internal quality of water milled glutinous rice flour.2.Aiming at the problem of adulteration of milled glutinous rice flour in real life market,this paper studies the problem of blending milled glutinous rice flour into milled glutinous rice flour.The NIR of milled glutinous rice flour samples were collected by Texas Instruments spectrometer,and the amylose content was detected by standard method.On this basis,the quantitative prediction model of amylose content was established by chemometrics method.The chemometric models based on joint interval algorithm(SI),non information variable elimination algorithm(UVE)and competitive adaptive reweighting sampling algorithm(CARS)combined with PLS were established respectively,and the models established by screening variables all had good results.Compared with joint interval partial least squares(SI),non information variable elimination algorithm(UVE)and competitive adaptive reweighting sampling algorithm(CARS),the NIR model established by UVE had better performance Competitive adaptive reweighted sampling algorithm(CARS)and all variable PLS algorithm get better results.The correlation coefficient of correction Rc= 0.9885,the correlation coefficient of prediction Rp = 0.9881,the corresponding root mean square error of correction set and prediction RMSEC = 1.4878,RMSEP= 1.6312.The results showed that the model of amylose content to detect whether the milled glutinous rice flour was mixed with rice flour had good effect.Whether the amylose content is less than 5% can be judged by the model.3.In view of the limitations of high price,professional operation and inconvenient carrying of near-infrared spectroscopy equipment on the market,a portable near-infrared spectroscopy detection system based on Android platform is developed in this chapter to realize the rapid onsite detection of the internal quality of water milled glutinous rice flour.Through Android Studio software platform,smart spec NIR software based on DLP NIR scan Nano micro NIR spectrometer was developed.On this basis,a Portable NIR rapid detection system was integrated.Then 15 independent water milled glutinous rice flour samples were randomly selected to verify the actual detection performance of the system.The results showed that the coefficient of variation(CV)was less than 1.08%,and the relative error(Re)was less than 2.61%;The coefficient of variation of three parallel detection results of acidity was CV < 5.63%,and the relative error was Re ? 8.69%.The results show that the portable detection system has good accuracy and stability in the detection of moisture and acidity of water milled glutinous rice flour,and its performance is superior.20 samples of water milled glutinous rice flour which have nothing to do with modeling were selected,and 10 of them were mixed with non quantitative rice flour.The amylose content of the samples was predicted three times by using the system.The results were compared with the actual value(standard value)of the national standard.The coefficient of variation CV of amylose was less than2.29%,and the accuracy was 100%,It completely meets the rapid detection standard of the internal quality of water milled glutinous rice flour.There was no significant difference between the predicted value and the standard value.The results show that the Portable NIR spectrum detection system developed in this study can realize the rapid on-site detection of the internal quality of water milled glutinous rice flour,and has certain detection accuracy and stability.In this study,water milled glutinous rice flour was taken as the research object,and the NIRS detection method for the internal quality of water milled glutinous rice flour was studied.The quantitative detection model of water content,acidity and amylose content of water milled glutinous rice flour with high accuracy and strong robustness was established to realize the rapid and non-destructive detection of the internal quality of water milled glutinous rice flour,In view of the phenomenon of rice flour mixed with water milled glutinous rice flour in the current market,the adulteration model of water milled glutinous rice flour was established.On this basis,a miniaturized,low-cost and practical portable detection system based on Android system is developed,and the actual detection performance of the system is verified,which is in line with the research expectation. |
PDF Full Text Request