Logo Lanfrica

Rebbouh-Mohamed/huawei-hakathon

Domaine:

environment and energyclimate

Type de record:

project
Créateur:
Reb
Hôte:
Multi-layer, real-time forest fire early detection and prevention in rural Algeria, combining IoT ground sensors, autonomous drones, satellite intelligence, and an AI-generated surveillance report engine — all deployed on Huawei Cloud using Docker containers orchestrated by Kubernetes. # 🔥 FireWatch AI — Forest Fire Prevention System ### AgriTech & Environmental Protection · AI, 5G & Digital Power > **Core focus:** Multi-layer, real-time forest fire **early detection and prevention** in rural Algeria, combining IoT ground sensors, autonomous drones, satellite intelligence, and an AI-generated surveillance report engine — all deployed on **Huawei Cloud** using Docker containers orchestrated by Kubernetes. --- ## 🌍 Problem Statement Algeria is one of the most wildfire-affected countries in the Mediterranean basin. Every summer, thousands of hectares of forest and farmland are devastated, threatening rural communities, biodiversity, and agricultural output. Traditional monitoring is **too slow** — fires are detected only after they've already spread. Our solution introduces a **three-layer, AI-driven detection system** that monitors fire risk at every scale: - **Layer 1 (before/)** — Ground IoT sensors detecting pre-ignition conditions - **Layer 2 (during/)** — Drones visually confirming fire with computer vision - **Layer 3 (after/)** — Satellites mapping active burns across the entire country All three layers feed into a **Gemini-powered intelligence report engine**, deployed and scaled on **Huawei Cloud**. --- ## 🏗️ Full System Architecture ``` ┌─────────────────────────────────────────────────────────────────────────────┐ │ DETECTION LAYERS │ │ │ │ 🌡️ LAYER 1 (before/) Ground Sensor Network │ │ IoT Sensors → MQTT → Ensemble AI (XGB+LGB+TabMLP+LR) → Supabase│ │ │ │ 🚁 LAYER 2 (during/) Drone Visual Detection │ │ Drone Camera → Flask API → YOLOv5 → Supabase │ │ …