🛡️ Smart AI Security System

An intelligent, Object-Oriented security camera system powered by Python, OpenCV, and Ultralytics YOLOv8. The system detects human presence in real-time, triggers external hardware alarms via Arduino, and automatically records video evidence.

Built with performance in mind, it utilizes multithreading to ensure a lag-free camera feed while the AI performs heavy object detection in the background.

---

🚀 Key Features

• Real-time AI Detection: Uses the yolov8n.pt model for fast, lightweight, and highly accurate human detection.
• Multithreading Architecture: AI inference runs on a dedicated background daemon thread, preventing UI blockage and maintaining high FPS.
• Hardware Integration: Communicates seamlessly with Arduino via Serial port to trigger physical alarms (e.g., LEDs, buzzers, sirens).
• Automated Evidence Collection: Automatically starts recording .avi video clips and saves .jpg snapshots the moment an intrusion is detected.
• Clean Code & OOP: Modular codebase divided into single-responsibility classes for easy maintenance, readability, and scalability.

---

🏗️ Project Architecture

The project was designed from scratch using Object-Oriented Programming (OOP) principles. Here is the breakdown of the core components:

1. Configuration - The central hub for all magic numbers and settings (ports, cooldowns, AI confidence thresholds).
2. Arduino - Hardware controller responsible strictly for Serial communication. Handles connection initialization, sending signals, and safe disconnections.
3. Recorder - Media manager that creates necessary directories (/recordings, /photos) and handles the video writing stream and snapshot generation.
4. Yolo - Implements the Producer-Consumer pattern. Runs on a separate thread, constantly processing frames provided by the main thread without blocking the camera feed.
5. SecuritySystem - The main Orchestrator. Connects all modules, captures the webcam feed, processes the business logic, and draws the HUD on the screen.

---

📋 Requirements

Hardware

• Standard Webcam (USB or built-in)
• Arduino board (Uno/Nano/Mega etc.) connected via USB

Software

• Python 3.9 or higher
• Required Python packages:
  • opencv-python
  • ultralytics
  • pyserial