Visible and near infrared (VIS-NIR) spectroscopy: measurement and prediction of soluble solid content of apple

Abstract

Both high nutrition and good tasting are the crucial factors of good agriculture product. Therefore, nondestructive measurement and prediction of fruit internal quality are an area that both technology and market section concern about. The objectives of this study were to evaluate the use of visible and near infrared (VIS-NIR) (380-1000 nm) spectroscopy for soluble solid content (SSC) prediction of apple; and to establish the relationship between the VIS-NIR reflectance spectra and the SSC of apples The reference SSC of apples was measured by using traditional destructive method. An artificial neural network with feedforward back-propagation (ANN-FFBP) was used to build the predictive model in this study. The ANN predictive model indicated good performance of SSC prediction with mean square error (MSE) of 0.1893 and a correlation coefficient (r) of 0.9668.