Fast and Non-Destructive Prediction of Moisture Content and Chologenic Acid of Intact Coffee Beans Using Near Infrared Reflectance Spectroscopy

Abstract

The main objective of this present research is to apply the near infrared reflectance spectroscopy (NIRS) as a fast and non-destructive method in predicting moisture content (MC) and chlorogenic acid (CGA) of intact roasted coffee beans. Diffuse reflectance spectrum were acquired for bulk coffee beans samples (Arabica and Robusta) in wavelength range from 1000 to 2500 nm. Spectra data were corrected and enhanced using standard normal variate (SNV). Prediction models, used to predict MC and CGA of intact coffee beans, were developed and performed using combination of principal component analysis (PCA) and multiple linear regression (PCA+MLR). The results showed that NIRS can be used to predict MC and CGA content of intact coffee beans simultaneously and rapidly with maximum coefficient correlation (r) were 0.92 for MC and 0.93 for CGA, whereas residual predictive deviation (RPD) indexed were 3.67 and 3.87 for MC and CGA content respectively. Based on the obtained results, it may conclude that near infrared spectroscopy can be applied in coffee quality evaluation especially to predict moisture content and chlorogenic acid. NIRS can be used and an alternative fast and non-destructive method bypassing standard laboratory procedures.