Non-Destructive Detection of Moisture Content Uniformity During the Drying Process of Maize by Hyperspectral Imaging Technology

Abstract:

Moisture content uniformity is one of the primary parameters in a drying process，which is important to evaluate the quality of dried-foods and the drying technique. The moisture content uniformity detected by the hyperspectral imaging technology was studied in the drying process of maize. A prediction model was developed by the mean and standard deviation features combined with the partial least squares(PLS)，where orthogonal signal correction method was used as the preprocessing method. The results showed that the prediction model developed by the mean and standard deviation features after preprocessing achieved the optimal performance with the correlation coefficient of 0.839 and the root mean square error of 1.74%，while the latent variables was reduced to 2 variables. Therefore，the hyperspectral imaging technology could be used as a non-destructive detection of moisture content uniformity.