🛡️ AtlasGuard

---

🚀 Overview

AtlasGuard is an AI-powered stadium security platform that combines computer vision and real-time analytics to detect violence, distress, and abnormal crowd behavior.
It leverages multimodal incident classification (video + audio) and integrates with modern web technologies to deliver a robust, scalable, and interactive solution for live stadium monitoring.

---

🏗️ Architecture & Features

🎥 Multi-Camera Monitoring

Grid View Dashboard: Live feeds from 4+ cameras simultaneously Real-time Processing: Sub-second incident detection Scalable Design: Easily expandable to additional camera feeds

🔊 Advanced AI Detection

Zero Manual Input: Fully automated incident classification Context-Aware Analysis: Considers crowd dynamics and venue layout Continuous Learning: Adapts to venue-specific patterns

📊 Interactive Dashboard

Real-time Notifications: Instant alerts with incident severity Historical Analytics: Trend analysis and reporting Custom Alert Rules: Configurable thresholds and responses

🛡️ Security Dispatch System

Stadium Layout Visualization: Interactive venue map Team Assignment: Direct dispatch to specific stadium sections Response Tracking: Monitor security team deployment and response times

Prerequisites

• Python 3.8+
• Node.js 16+
• OpenAI API key
• Camera feeds (IP cameras or video files for testing)

---

Installation

1. Clone the repository

   git clone [https://github.com/your-username/SafeArena.git](https://github.com/your-username/SafeArena.git)
   cd SafeArena

2. Set up the backend (FastAPI)

   cd backend
   pip install -r requirements.txt

3. Set up the frontend (React + Tailwind)

   cd frontend
   npm install

---

Configuration

1. Create .env file in the backend directory

   cp .env.example .env

2. Add your API key and camera streams

   OPENAI_API_KEY=your_api_key_here
   CAMERA_STREAMS=["rtsp://camera1", "rtsp://camera2", "rtsp://camera3", "rtsp://camera4"]
   DATABASE_URL=sqlite:///./safearena.db

---

Running the Application

1. Backend

   cd backend
   uvicorn main:app --reload --port 8000

2. Frontend

   cd frontend
   npm start

---

Accessing the Dashboard

Open your browser and navigate to: http://localhost:3000 Open your browser and navigate to: http://localhost:3000

Live Monitoring Interface

• 4-Camera Grid View: Simultaneous monitoring of multiple angles
• Real-time Status Indicators: Visual feed health and AI processing status
• Incident Overlay: On-screen alerts with bounding boxes around detected incidents

---

Alert Management

• Popup Notifications: Immediate incident alerts with severity levels
• Alert Queue: Manage and prioritize multiple simultaneous incidents
• Historical Log: Complete incident history with timestamps and classifications

---

Dispatch Control

• Interactive Stadium Map: Click-to-dispatch security teams
• Lane Assignment: Assign teams to specific stadium sections or lanes
• Response Tracking: Monitor team location and response status

👥 Authors

Aymane ELBEKKALI Master's Student in Advanced ML 📧 elbekkaliaymane@gmail.com 🔗 Linkedln

souhaib benbouazza phd student QML 📧 your.email@example.com 🔗 Linkedln

Fedy Ben Hassouna Your Title or Role 📧 your.email@example.com 🔗 Linkedln

Omar BOURJA Team Manager UM6P 📧 your.email@example.com 🔗 Linkedln