Research On Detection Technology Of Azoformamide Content In Flour Based On NIR Technology

There are so many food safety incidents in recent years, the flour as the basic material of a large amount of food, the additions contents are noted by more and more people, in order to monitor the additions in food timely, a fast and convenient method is needed. Azodicarbonamide as a flour gluten fortifier is wildly used in flour process in many countries, but it was proved by some researches to be dangerous or unhealthy for people and not suitable to be added in flour, hence, there is a need to identify the concentration of azodicarbonamide in flour. This article proposed a method of prediction to the content of azodicarbonamide in flour using near infrared spectroscopy technique based on physical characteristic analysis. Spectral data were obtained by NIRS DS2500 scanning 101 samples, Mahalanobis distance method and leave-one-out cross-validation method were used to eliminate abnormal spectral data, the derivative method, standard normal variate transformation, multiplicative scatter correction, wavelt transform and Savitzky-Golay smoothing were used to preprocessing data, correlation coefficient method, dominance transformation analysis method and principal component analysis method were used to choose characteristic wave bands. Partial least squares, back propagation neural network and radial basis function are used to establish predictive model. By comparing these three model, radial basis function model has the best predictive results and the correlation coefficients, root mean square error of prediction and relative percent deviation of the optimized radial basis function model reached 0.9992, 3.4007 and 25.0350.