A Performance Evaluation of Vis/NIR Hyperspectral Imaging to Predict Curcumin Concentration in Fresh Turmeric Rhizomes

Hyperspectral image (HSI) analysis has the potential to estimate organic compounds in plants and foods. Curcumin is an important compound used treat a range of medical conditions. Therefore, method rapidly determine rhizomes with high curcumin content on-farm would be significant advantage for farmers. varies within, between varieties but current chemical methods are expensive time consuming. This study compared three turmeric (Curcuma longa) examined laboratory-based HSI predict using visible–near infrared (400–1000 nm) spectrum. images (n = 152) fresh rhizome outer-skin flesh were captured, local (yellow, orange, red). Distribution curcuminoids total was analysed. Partial least squares regression (PLSR) models developed concentrations. Total proportion differed significantly among all varieties. Red had highest concentration (0.83 ± 0.21%) orange (0.37 0.12%) yellow (0.02 0.02%). PLSR predicted raw spectra pooled data (R2c 0.83, R2p 0.55, ratio prediction deviation (RPD) 1.51) slightly improved by single variety (orange) only 0.85, 0.62, RPD 1.65). However, poor 0.64, 0.37, 1.28). These can discriminate ‘low’ ‘high’ values so may adapted into two-level grading system. help identify concentrations allow more efficient refinement medicinal purposes.