Prediction of sugar content along the height of sugarcane internodes using VisNIR

Abstract

Sugarcane is a tall crop with sugar distribution varying along the height. Using Vis/NIR spectrometer, an initial measurement was carried out to investigate the sugar distribution along the height of internodes. Spectral data was collected from 25 internode samples. Each internode sample was further cut into four sections and scanned based on cross-sectional scanning (CSS) and skin scanning (SS). It was found that for the cross-sectional scanning (CSS), the coefficient of determination (R2) varied between 0.78 and 0.89, achieving the highest value around the middle of the internode. For the skin scanning (SS), the coefficient of determination (R2) varied between 0.81 and 0.91, also achieving the highest value around the middle of the internode. The results suggested that quality prediction based on non-juice sample is feasible using a spectrometer.