Ripeness stage characterization of coffee fruits (coffea arabica L. var. Castillo) applying chromaticity maps obtained from digital images

Abstract

Abstract The harvesting is one of the most expensive stages during Colombian coffee production since the selectivity of ripe fruits (from the trees) is carried manually nowadays. However, the scarcity of workforce has become a new challenge for developing technologies that can be used in situations that traditional mechanized harvesters cannot be implemented. The classification of coffee fruits through the ripeness could help create selective harvesting methods from the different engineering fields (robotic, artificial vision, vibratory systems, etc.). This study describes a methodology to characterize the coffee fruit ripeness (Coffea arabica var. Castillo) using color maps obtained from digital image processing and colorimetry. Ninety-two samples were measured, and for each sample, a digital image was processed to extract the most representative color with an algorithm (RGB space), and parallel to this, the chromaticity was measured with a colorimeter (CIELab space). Both color spaces were related through linear transformations computed from the data. Results show that the ripeness classification is well defined in both color spaces; however, RGB space was divided into four ripening since the ripeness timeline was linearly represented. Through the transformations, the image data were contrasted with those measured with the colorimeter, which demonstrated that it could be converted in a colorimeter under certain specific conditions.