Estimation of a Color Uniformity Index in Mandarins Using an Artificial Intelligence-Based System

In the agricultural industry, color uniformity is currently considered a key criterion for determining fruit quality and commercial acceptability, especially in citrus, where color uniformity is particularly relevant, as it is closely related to the degree of ripeness, perceived freshness and overall product attractiveness. Among citrus fruits such as lemon, mandarin, grapefruit, orange, mandarin stands out as an important crop due to its high demand and consumer sensitivity to color variations, which makes uniformity in appearance directly influence buyer perception and compliance with trade standards. However, the evaluation of color uniformity is still done manually, which introduces subjectivity, errors and inconsistency in grading processes, negatively impacting the efficiency and competitiveness of producers. Several studies have explored the analysis of citrus using artificial intelligence. However, many of these efforts face limitations related to image quality, lack of standardization in capture conditions, or limited adaptation to local contexts such as the Colombian market. To address these challenges, an artificial intelligence-based system was developed to quantitatively assess color uniformity in postharvest mandarins. The system incor-porates an image acquisition subsystem, followed by detection, segmentation and data analysis procedures to calculate a color uniformity index, with the objective of improving the selection and classification processes in the citrus industry within the Colombian sector. The use of an artificial intelligence algorithm and the definition of a quantitative index of color uniformity in mandarin oranges is proposed in order to provide agricultural producers with an objective and accurate tool to optimize the classification of their fruits.