MBONOMA/Agrihub

AgriHub: A full-stack web and mobile application for modern agriculture management. Features include AI-powered crop recommendations, real-time market data, multilingual support (English & Kinyarwanda), and interactive dashboards. Built with a responsive frontend, serverless backend, and Firebase/Realtime Database integration. # AgriHub Offline plant disease detection mobile app powered by on-device machine learning. Identify crop diseases from leaf photos without an internet connection. ## Features - **Offline ML Inference** -- Runs MobileNetV2 on-device using ONNX Runtime, no server required - **38 Disease Classes** -- Covers 14 crop species including Apple, Tomato, Corn, Grape, Potato, and more - **Camera & Gallery** -- Capture a leaf photo or pick from gallery for instant analysis - **Disease Details** -- View confidence scores, severity, and crop identification - **Scan History** -- Track past scans with local storage - **Treatment Info** -- Organic and chemical treatment recommendations ## Tech Stack | Layer | Technology | |-------|-----------| | Framework | React Native 0.81 + Expo SDK 54 | | Navigation | Expo Router 6 (file-based) | | ML Runtime | ONNX Runtime React Native 1.23 | | ML Model | MobileNetV2 (PlantVillage dataset) | | Image Processing | expo-image-manipulator + jpeg-js | | Storage | AsyncStorage | | Language | TypeScript 5.9 | ## Project Structure ``` AgriHub/ ├── app/ # Expo Router screens │ ├── (tabs)/ # Tab navigation │ │ ├── index.tsx # Home / scan screen │ │ ├── browse.tsx # Browse crops & diseases │ │ ├── history.tsx # Scan history │ │ └── settings.tsx # App settings │ ├── scan/ │ │ ├── camera.tsx # Camera capture │ │ └── results.tsx # Analysis results │ └── details/ │ ├── crop/[id].tsx # Crop detail page │ └── disease/[id].tsx # Disease detail page ├── src/ │ ├── ml/ # Machine learning │ │ ├── ModelManager.ts # ONNX inference engine │ │ ├── labels.ts # 38 disease + 14 crop labels │ │ └── types.ts # ML type definitions │ ├── components/ui/ # Reusable UI components │ ├── constants/ # App & ML config │ ├── context/ # R …