Logo Lanfrica

brytesika-AI/AgroResilience-Africa---Continental-Crop-Monitoring-AI-Forecasting-System

Domain:

agriculturegeospatial

Record type:

softwareproject
Creator:
bry
Host:
# 🌍 AgroResilience Africa: Continental Crop Monitoring & AI Forecasting System An enterprise-grade, edge-computed geospatial platform designed to track crop health, estimate yields, and forecast agricultural productivity across Africa. Powered by **Digital Earth Africa's Open Data Cube (STAC API)** and deployed entirely on the **Cloudflare Edge Network** with integrated **Workers AI**. ## πŸš€ Key Capabilities - **Continental Scaling with Granular Filters:** Seamless navigation from a pan-African overview down to specific national fields (e.g., Zambia's Mkushi Farm Block). - **Edge Geospatial Compute:** Real-time computation of the **Enhanced Vegetation Index (EVI)** directly inside Cloudflare Workers using Sentinel-2 Surface Reflectance data, completely bypassing heavy GIS server overhead. - **Robust Cloud Masking:** Automated processing of the Scene Classification Layer (SCL) to remove cloud interference, matching real-world data science practices. - **Edge AI Insights:** Utilizes Cloudflare Workers AI (`Llama-3`) to run inference on time-series vegetation analytics to produce descriptive yield forecasts and strategic resilience recommendations. ## πŸ› οΈ Architecture & Tech Stack ```text [User Interface: React + Tailwind + Leaflet] β”‚ (JSON API via HTTPS) β–Ό [Edge Layer: Cloudflare Workers + Pages] β”‚ β”œβ”€β–Ί [AI Inference: Cloudflare Workers AI (Llama-3)] β”œβ”€β–Ί [Cache / Metadata: Cloudflare KV Store] └─► [Geospatial Core: Digital Earth Africa STAC API] ``` - **Frontend:** React 18, TypeScript, Tailwind CSS, Vite, Recharts, Leaflet. - **Backend:** Cloudflare Pages Functions, Hono Framework. - **Geospatial Engine:** Digital Earth Africa STAC API & AWS Sentinel-2 Level-2A Catalogs. - **Artificial Intelligence:** Cloudflare Workers AI Engine (`@cf/meta/llama-3-8b-instruct`). - **Database/Caching:** Cloudflare KV Namespace (`CROP_CACHE`). ## πŸ“Š Grounded Regional Use Cases (Zambia & Tanzania) Inspired by real-world Earth Observation initiatives: 1. **Mkushi Farm Block (Zambia) …