Smartphone Crop Disease Detection for Cassava

Overview

Cassava, a staple food for millions of Ugandans and a cornerstone of regional food security, is under constant threat from devastating diseases that can wipe out entire fields before a farmer even knows something is wrong. Yet access to agricultural expertise in rural Uganda is scarce, expensive, and often arrives too late.

Smallholder cassava farmers in Uganda face a critical surveillance gap: crop diseases often take hold silently, spreading through planting materials and across communities long before visible symptoms appear. Without affordable, accessible diagnostic tools, farmers rely on guesswork or delayed expert advice, by which time significant damage is already done. Compounding this, rural farming communities often lack reliable internet connectivity and the financial means to access specialised agricultural services.

The project deploys machine learning models on low-end smartphones enabling farmers to detect disease at the earliest stages, often before symptoms are even visible to the naked eye. Low-cost 3D-printed hardware keeps the technology affordable, while AI-guided guidance helps farmers identify and select clean planting materials to prevent future infection and break cycles of reinfection. Crucially, the team co-designed the solution directly with farming communities, ensuring the tools reflect real needs, fit existing practices, and are genuinely adoptable at scale.

The result is a scalable, community-grounded model for AI deployment in low-resource agricultural settings. By catching disease early and supporting better planting decisions.