Pendugaan kadar air, protein dan karbohidrat biji sorgum secara non-destruktif dengan metode near infrared (NIR)
Abstract
Sorghum is local resources that can be used as a substitution of corn on feed. Moisture, protein and carbohydrate contents of sorghum can be predicted using near infrared method. The aim of this research was to apply NIR method to analyze moisture, protein and carbohydrate contents of sorghum. The calibration methods that used in this research are principal component regression (PCR) and partial least square (PLS). Some data treatments are applied on absorbance and reflectance spectrum. They are smooth average 3 points, normalization between 0-1, second derivative Savitzky-Golay 9 points, combination of smooth average 3 points and second derivative Savitzky-Golay 9 points, and combination of smooth average 3 points, normalization between 0-1, and second derivative Savitzky-Golay 9 points. PLS method with absorbance data and combination of smooth average 3 points, normalization between 0-1, and second derivative Savitzky-Golay 9 points is the best calibration method and data treatment to predict moisture content of sorghum. The best calibration method to predict protein content is PLS with absorbance data and second derivative Savitzky-Golay 9 points data treatment. PLS method with absorbance data and second derivative Savitzky-Golay is the best calibration method and data treatment to predict carbohydrate content. Standard error of prediction (SEP) and coefficient of variability (CV) respectively were 0.005% and 0.04% for moisture content prediction, 0.01% and 0.18% for protein content prediction, 0.01% and 0.02% for carbohydrate content prediction.