Research On Quality Inspection Of Cookies Based On Multi - Spectral Imaging Technology

Cookie as a fast-moving consumer good, appreciated for its unique flavor and taste, attracts the consumers. As cookies have been in large demand, some fraudulent phenomena have increased, these behaviors harm to consumers. In order to guarantee the order of market economy, safety control and quality evaluation are the necessary procedures. The traditional quality and safety controls for analyses are mainly through chemical analysis and high performance liquid chromatography (HPLC). However, these methods are destructive and environmental damage. Therefore, it is urgently required to establish a rapid and non-destructive approach for quality evaluation of cookies. Based on this situation, this paper aims at developing a rapid and thoroughly determination method for quality evaluation of cookies using multispectral imaging technology system (405-970nm) with chemometrics methods-partial least square (PLS) and back propagation neural network (BP-ANN). The major research contents and conclusions are as follows:(1) Multispectral imaging system was used to determine the borax in both untreated flour and baking products. The results shew that the BP-ANN method had a better prediction performance than the PLSR method in untreated flours experiment, as the correlation coefficients of determination in predicted test Rp2 and standard error of prediction (SEP) acquired by the BP-ANN method were 0.821 and 134.269; In the experiment of baking products, PLSR had a better prediction performance than the BP-ANN method, the correlation coefficients of determination in predicted test Rp2 using PLSR model were 0.842 with SEP was 110.75.(2) Studied the application prospect of multispectral imaging technology in the determination of the rancidity in butter cookies. In the present paper, the potential of applying multispectral imaging technology with chemometrics methods (PLSR and BP-ANN) to evaluate the moisture content (MC), acid value (AV) and peroxide value (PV) in butter cookies was investigated. The optimal model for predicting MC, AV and PV were obtained by PLSR with correlation coefficient (r) of 0.909,0.944 and 0.971.