Non-destructive and rapid measurement of sugar content in growing cane stalks for breeding programmes using visible-near infrared spectroscopy

Abstract

The efficient selection of sugarcane varieties with enhanced sugar content requires simple, rapid, accurate and cost-effective assays. The objective of this research was to develop spectroscopic models for non-destructive evaluation of Commercial Cane Sugar (CCS) in growing cane stalks. A portable visible-shortwave near-infrared (Vis/SWNIR) spectrometer with a wavelength range of 570–1031 nm was applied to cane stalks grown under normal field conditions. The CCS models were developed by partial least squares (PLS) regression using spectra sets obtained at three different times (i.e. morning, afternoon and evening), based on both individual spectra and a combined set. During the in-field measurements, it was found that model performance could be affected by varying sample and instrument temperatures, especially for the combined set. The models had coefficients of determination of the prediction set (r2) of 0.76, 0.76, 0.78 and 0.69 and root mean square errors of prediction (RMSEP) of 1.01, 1.05, 0.99 and 1.17 CCS for the models constructed with the spectra sets of the morning, afternoon, evening and combined, respectively. These results indicate that the CCS models could be used for the monitoring and screening of sugarcane clone selection, except for the model obtained from the combined spectra set, which was influenced by varying instrument temperature. Thus, heat protection for the portable Vis/SWNIR instrument, allowing it to maintain a constant instrument temperature, should be considered if data collection across different periods of the day is necessary.