ade-daemon/ai-crop-doctor

Offline-first AI crop disease diagnosis app for Nigerian smallholder farmers. Snap a photo, get instant treatment advice in Yoruba, Hausa, Igbo, or Pidgin. Built with FastAPI, React Native, and TensorFlow Lite. # 🌿 AI Crop Doctor **Offline-capable, multilingual AI crop disease diagnosis and advisory for Nigerian smallholder farmers.** AI Crop Doctor uses on-device machine learning to diagnose crop diseases and pest infestations from smartphone photos, then provides treatment recommendations tailored to locally available inputs. Built on a disease image dataset collected from Nigerian farms. --- ## 🎯 Problem - Nigerian smallholder farmers lose **40-50% of crop yields** annually to pests and diseases - Fewer than **5,000 extension agents** serve over **70 million farming households** - Existing AI crop tools (Plantix, etc.) are trained on non-African datasets with limited accuracy on Nigerian crop varieties - Most rural areas have **intermittent or no internet connectivity** ## 💡 Solution A mobile app that: - **Diagnoses crop diseases** from phone camera photos using on-device ML (works offline) - Covers **15 Nigerian staple and cash crops** (cassava, maize, rice, yam, cowpea, tomato, pepper, plantain, cocoa, soybean, sorghum, millet, groundnut, melon, okra) - Provides **treatment recommendations** using locally available inputs - Supports **voice interaction in Yoruba, Hausa, Igbo, and Pidgin English** - Creates **geo-tagged outbreak maps** for agricultural authorities ## 🏗️ Architecture ``` ┌─────────────────────────────────────────────────┐ │ Mobile App │ │ (React Native + TFLite) │ │ │ │ ┌──────────┐ ┌──────────┐ ┌───────────────┐ │ │ │ Camera │ │ Offline │ │ Voice AI │ │ │ │ Diagnosis │ │ ML │ │ (Yoruba/Hausa │ │ │ │ │ │ Inference│ │ Igbo/Pidgin) │ │ │ └──────────┘ └──────────┘ └───────────────┘ │ │ │ │ │ ┌───────┴────────┐ │ │ │ Local Storage │ │ │ │ (SQLite/MMKV) │ …