dheerajpatel3448w/Ecowatch-geo-intelligence

# 🌍 EcoWatch: Geospatial Intelligence & Threat Detection System **EcoWatch** is an advanced satellite threat detection platform built to combat deforestation, illegal mining, and ecological damage. It combines a Next.js command-center UI, Apache Kafka for high-throughput event streaming, and Python-based Machine Learning (Qwen2-VL) to analyze satellite imagery in real-time. ## 🚀 Key Features - **🌐 Global Threat Radar:** A 3D geospatial dashboard with real-time alerts, metrics, and threat mapping. - **🛰️ Satellite AI Analysis:** Integrates the Qwen2-VL-2B-Instruct Vision-Language Model to automatically detect deforestation and environmental anomalies from satellite imagery (NDVI analysis). - **📡 Mission Control (Zones):** Draw and manage surveillance zones directly on a map using React-Leaflet. Trigger manual or automated ML scans for specific coordinates. - **⚖️ Legal & Carbon Engine:** Calculates exact Carbon loss and generates official **PDF Legal Evidence Reports** automatically for FIR submission. - **📱 Field Operations Portal:** A mobile-first UI for ground rangers to upload evidence photos, matched with GPS locations, which are also verified locally by AI. - **⚡ Real-Time Streaming:** End-to-end Kafka integration handling heavy ML workloads asynchronously between Node.js and Python. ## 🏗️ Architecture & Tech Stack The platform is divided into three main microservices, heavily utilizing containerization for the infrastructure. ### 1. Frontend (`/ecowatch`) - **Framework:** Next.js (App Router), React 19 - **Styling:** Tailwind CSS v4, Framer Motion (Military/Command-Center Dark Theme, Glassmorphism) - **Mapping & 3D:** React-Leaflet, Mapbox-GL, Three.js, React-Three-Fiber - **Data Visualization:** Recharts - **Real-Time:** Socket.IO Client ### 2. API Backend (`/node-service`) - **Runtime:** Node.js, Express.js - **Database:** MongoDB (Mongoose) - **Message Broker:** KafkaJS (Producer/Consumer) - **Features:** JWT Authentication, Role-Based Access Control (Admin, Analyst, Ranger), PDFKit (Legal Reports), Multer (Image Uploads) ### 3. ML Service (`/ml-service`) - **Runtime:** Python 3.11+, FastAPI - **AI Models:** Qwen2-VL-2B-Instruct, HuggingFace Transformers, PyTorch - **Message Broker:** Confluent-Kafka - **Features:** Satellite image processing, NDVI physics logic, local photo validation ### 4. Infrastructure (`docker-compose.yml`) - **Zookeeper & Kafka:** Confluent images for event streaming. - **MongoDB:** Database storage. ## 📂 Project Structure ecowatch/ ├── ecowatch/ # Next.js Frontend UI ├── node-service/ # Express API Backend ├── ml-service/ # FastAPI Python Machine Learning Service ├── docker-compose.yml # Infrastructure configuration (Kafka, Mongo) ├── ECOWATCH_HANDOFF.md # Developer architectural notes └── README.md # Project documentation ## ⚙️ Setup & Installation ### 1. Start Infrastructure (Docker) Ensure Docker Desktop is running, then spin up MongoDB, Zookeeper, and Kafka. docker-compose up -d ### 2. Setup Node.js API Service cd node-service npm install npm run dev *(Runs on `http://localhost:5000`)* ### 3. Setup Python ML Service *Note: Run locally to utilize system GPU or CPU for heavy ML processing.* cd ml-service python -m venv .venv # Activate venv: .venv\Scripts\activate (Windows) OR source .venv/bin/activate (Mac/Linux) pip install -r requirements.txt uvicorn api.app:app --host 0.0.0.0 --port 8001 *(Runs on `http://localhost:8001`)* ### 4. Setup Next.js Frontend cd ecowatch npm install npm run dev *(Runs on `http://localhost:3000`)* ## 🔐 Environment Variables You will need `.env` files in each respective directory. Here is a basic template based on the architecture: **`/node-service/.env`** PORT=5000 MONGO_URI=mongodb://localhost:27017/ecowatch JWT_SECRET=ecowatch_production_secret_change_me KAFKA_BROKER=localhost:9092 **`/ml-service/.env`** KAFKA_BROKER=localhost:9092 KAFKA_GROUP=ml-workers SH_CLIENT_ID=your_sentinel_hub_client_id SH_CLIENT_SECRET=your_sentinel_hub_client_secret HF_TOKEN=your_huggingface_token MODEL_NAME=Qwen/Qwen2-VL-2B-Instruct LOG_LEVEL=INFO ## 👨‍💻 Developer & Maintainer Built by **Dheeraj Patel** (MERN Stack & AI Developer).