Diseño de una metodología experimental para la discriminación de masas del Café Esmeralda durante su proceso de clasificación y selección

Abstract

This research presents the development of an objective experimental methodology for the classification, selection, and mass estimation of Esmeralda coffee beans, a critical need in the specialty coffee industry.The approach integrates digital colorimetry in the CIE L∗a∗b∗ space with morphological image analysis. Defective beans (e.g., broca-affected, stained) are discriminated using sequential filters based on empirical thresholds defined in the chromatic channels, validated by the clear separation between healthy and diseased beans in the L∗a∗b∗ space.A robust linear relationship was established between the projected area A (mm2) and mass m (g) (A = 31240.9 · m − 1871.2, R2 = 0.8261). Although the relative uncertainty per individual bean is considerable (≈ 10.6%) due to inherent morphological variability, the method demonstrated high accuracy in aggregate batch valuation, achieving a relative error of 1.93% in total mass estimation.The methodology is encapsulated in an interactive graphical interface, providing a quantifiable and traceable tool for quality control in coffee processing.