Objective:
Develop an AI-powered application that automatically sorts coffee cherries based on color, size, and density. This would improve the consistency of premium quality and enhance efficiency in the coffee processing chain.
Context:
In Kenya’s 2022/2023 coffee season, 48,649 metric tons of clean coffee were produced, valued at USD 204.5 million, directly or indirectly supporting 6 million Kenyans. Sorting coffee cherries is a critical step in ensuring the quality of the final product. Manual sorting, which removes defective, underripe, overripe, and diseasedcherries, is time-consuming and labor-intensive.
Additionally, density sorting, where cherries are submerged in water to remove underdeveloped ones, further separates low-quality cherries but needs to be complemented with more advanced sorting technologies.
Sorting by color alone can be subjective, particularly for yellow cherries, and manual processes lead to inefficiencies and misclassifications, affecting quality and pricing. Furthermore, sorting (which occurs at the cherry level) is different from grading (which occurs at the green bean level), where beans are categorized into AA, AB, PB, C, E, TT, and T grades. The proposed solution aims to automate cherry sorting to ensure superior quality control, increase efficiency, and provide early feedback on coffee quality. This will benefit farmers, cooperatives, and the broader coffee supply chain. This will support economic growth (SDG 8) and promote innovation in the coffee industry (SDG 9).
Expected outcomes:
AI-powered application that sorts coffee cherries based on color, size, and density, capable of removing defective, diseased, underripe, and overripe cherries. Real-time quality feedback for farmers and cooperatives, allowing them to categorize cherries early in the processing chain for better market positioning. Scalable solution that smallholder farmers and large cooperatives can adapt to enhance the efficiency and accuracy of cherry sorting. Enhanced coffee quality at the processing stage, leading to more consistent and higher grading outcomes during green coffee processing.