A rapid determination of wheat flour’s components based on near infrared spectroscopy and chemometrics

Abstract

Abstract In this work, a rapid and simple analytical method for the quantitative determination of moisture, protein, wet gluten, starch and sedimentation index in the wheat flour was established by the combination of near infrared spectroscopy and chemometrics. The spectra of the 229 wheat flour samples were collected by a portable near infrared fast analyzer. The contents of these components were determined according to the relevant Chinese National Standards, and were taken as the corresponding reference database. Partial least squares (PLS) modeling method was used, and seven spectral pretreatment methods were employed to eliminate the optical interference from background and other noise information, then the quantitative analysis effect of the model on wheat flour index was investigated. The best result was obtained with FD + SG(15, 3) + MC method for moisture, protein, wet gluten and sedimentation index, FD + SG(15, 2) + MC method was more suitable for starch. The principal component numbers (PCs) were also optimized to obtain a superior model effect. In this study, the predicted values and the reference values in the PLS model of wheat flour showed superior linear relationship within the calibration range. Moreover, the absolute error of the predicted values and their corresponding reference values were within their confidence intervals, respectively. The deviations for moisture, protein, wet gluten and starch fluctuated less, only sedimentation index fluctuated greatly. The actual prediction correct rate of moisture, protein, wet gluten, starch and sedimentation index were 96.8%, 96.8%, 90.3%, 100.0% and 80.6%, respectively, which indicated the prediction was excellent.