Intrusion Detection System with Real Time Telegram Alerts

A computer vision–based intrusion detection system that uses YOLOv5 for person detection, OpenCV for Masking, and a chatbot for sending real time alerts to Telegram. The added functionality is that the user can define his Region of Interest (ROI) by drawing a ploygon on the feed and detection happens in that ROI only. A confidence threshold of > 60% is used along with debounce and false-positive control which leads to 98% accuracy in alert triggering. An alert message acoompanied by a captured frame of intrusion is then triggered to the user's Telegram.

This project can be practically deployed for intrusion detection on campus frames and boundaries and can be extended to other industrial usage like PPE detection in a particular ROI.

System Architecture:

flowchart TD
    A[Video Stream Input] --> B[ROI Selection<br/>Polygon Mask]
    B --> C[YOLOv5 Person Detection]
    C --> D[Confidence Threshold Filtering]
    D --> E[Temporal Window Logic]
    E --> F[Telegram Alert Trigger]

Loading

Accuracy & False Positive Control

Intrusion alerts are triggered only when all conditions are satisfied:

• A person is detected inside the ROI
• Detection remains stable across multiple frames
• Minimum positive frame count is reached
• Cooldown period since last alert has elapsed

This approach significantly reduces:

• One-frame false positives
• Flickering detections
• Alert flooding

Achieved ~98% reliable alert triggering during controlled testing

Demo Screenshots

ROI Polygon Selection

Detection Within Masked ROI

Telegram Alert With Captured Frame

Installation and Setup

1. Clone the repository:

   git clone https://github.com/Aman-Sarin/intrusion-detection-telegram-alert.git
   cd intrusion-detection-telegram-alert

2. Create virtual environment:

   python -m venv venv
   source venv/bin/activate   # Linux/Mac
   venv\Scripts\activate      # Windows

3. Install dependencies:

   pip install -r requirements.txt

Telegram Bot Setup

Create a bot using BotFather Get:

BOT_TOKEN
CHAT_ID

Update them in main.py in place of

import os
    BOT_TOKEN = os.getenv("TELEGRAM_BOT_TOKEN")
CHAT_ID = 12345678

Running the project

python src/main.py

Controls:

1. Press q to exit video stream
2. Selection of polygon with left mouse button and finalizing ROI with right mouse button

Future Improvements

1. Send intrusion video alerts
2. Multi-camera support
3. Person tracking (DeepSORT)
4. Other use cases apart from human - PPE, any animal, etc.

Author

Aman Sarin Assistant Manager – Supply Chain (Industrial Digital Transformation) | Electronics & Instrumentation Engineer (NIT Rourkela)

🔗 GitHub: https://github.com/Aman-Sarin | 🔗 Interests: Computer Vision, Industrial Automation