🛡️ AI Surveillance Pro: Airport Security & Analytics System

This project is a comprehensive end-to-end security solution designed for high-traffic areas like airports and customs zones. It leverages real-time computer vision and artificial intelligence to detect border breaches and track specific entities.

Unlike standard object detection, this system utilizes Instance Segmentation to precisely mask objects, providing granular data analysis and high-fidelity visual evidence.

🚀 Key Features

• 🔍 AI Engine: Powered by YOLOv8-Segmentation for real-time tracking of passengers and "suspicious" items (bottles, bags, etc.).
• ⚡ Hardware Acceleration: Fully optimized for Apple Silicon (M1/M2/M3) using MPS (Metal Performance Shaders) for high-FPS processing.
• 📊 Live Dashboard: A sleek, web-based interface displaying real-time crossing statistics and captured violation snapshots.
• 🗄️ Persistent Storage: Integrated with a PostgreSQL database to log every breach with timestamps, unique object IDs, and image paths.
• 🛠️ Modern Architecture: Built with FastAPI (Python) for asynchronous data handling and low-latency API communication.

🛠️ Technology Stack

Component	Technology
AI & Vision	Python, Ultralytics YOLOv8-Seg, OpenCV
Backend	FastAPI, Uvicorn
Database	PostgreSQL, Psycopg2
Frontend	HTML5, Bootstrap 5, JavaScript (Fetch API)
Security	Python-Dotenv (Environment Variable Management)

📦 Installation & Setup

1. Clone the Repository:

   git clone [https://github.com/euzghe/AI_Surveillance_Pro.git](https://github.com/euzghe/AI_Surveillance_Pro.git)
   cd AI_Surveillance_Pro

2. Environment Setup:

   python3 -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt

3. Database Configuration: Create a .env file in the root directory and define your credentials:

   DB_PASSWORD=your_secure_password
   

4. Run the System:

   • Start Backend: uvicorn api:app --reload
   • Start AI Engine: python main.py
   • View Dashboard: Open index.html in your browser.

🎯 Project Logic (Vertical Border Breach)

The system overlays a virtual vertical line on the video feed. When an entity crosses this line:

1. The AI segments (masks) the object for visual evidence.
2. A high-resolution snapshot is saved locally.
3. The metadata is logged into the PostgreSQL database.
4. The Web Dashboard UI updates instantly via polling the API.

---

⭐ Developed as a Senior Project for Software Engineering (4th Year).