AI Surveillance System

Overview

The AI Surveillance System is an object detection and tracking application that uses YOLO (You Only Look Once) to detect people in a video feed and trigger alerts for suspicious activity such as loitering. The system is built using OpenCV, NumPy, and Pygame for real-time video processing and alert sound playback.

Features

• Real-time person detection using YOLOv3-tiny.
• Tracks movement of detected persons.
• Detects loitering behavior and triggers an audio alert.
• Auto-stop feature to terminate after a set duration.
• Graceful exit with 'q' key press or keyboard interruption.

Installation

Prerequisites

Ensure you have Python 3.x installed, then install the required dependencies:

pip install opencv-python numpy pygame

Download YOLO Weights & Config Files

1. Download YOLOv3-tiny files:
   • yolov3-tiny.weights
   • yolov3-tiny.cfg
   • coco.names
2. Place these files inside a yolo/ directory in the project folder.

Usage

Run the script to start the AI surveillance system:

python ai_surveillance.py

Controls

• Press 'q' to quit the application.
• The system will auto-stop after a set duration (default: 60 seconds).

Configuration

You can as well adjust parameters in the script:

Parameter	Description	Default Value
loitering_threshold	Frames before loitering alert	100
auto_stop_time	Auto-stop time (seconds)	60
alert_sound	Path to alert sound file	"alert.mp3"

Project Structure

├── yolo/
│   ├── yolov3-tiny.weights
│   ├── yolov3-tiny.cfg
│   ├── coco.names
├── alert.mp3
├── ai_surveillance.py
|__ sv.py
├── README.md

License

This project is licensed under the MIT License.

Credits

• YOLOv3: Joseph Redmon & Ali Farhadi
• OpenCV: OpenCV.org

---

🚀 Developed for AI-powered surveillance & security monitoring.