A real-time system that combines weapon detection with automated alerts and data logging. Ideal for enhancing public safety in high-risk environments like campuses, malls, and transit hubs.
The increasing number of public threats involving weapons necessitates proactive detection mechanisms. ARMDET provides an integrated approach to detecting harmful objects (e.g., knives, guns) in visual feeds and instantly alerting authorities with precise location and timestamp metadata.
The system is composed of:
- YOLOv5-based Detection Engine
- Flask-based API for prediction and response
- Streamlit UI for interactive usage
- Twilio-based SMS Alert System
- Firebase Storage and Realtime Database
git clone https://github.com/yourusername/ARMDET.git
cd ARMDETpip install -r requirements.txtMake sure
api.pyis correctly set up with model paths and Twilio keys
Mac/Linux:
export FLASK_APP=api.py
flask runOpen a new terminal and run:
streamlit run app.py| Mode | Description |
|---|---|
| Image Upload | Detects weapons from uploaded image files |
| Video Upload | Parses video frames and flags potential threats |
| Live Stream | Real-time webcam feed detection with automated alerting |
| Check Logs | Visual interface to view historical detections and metadata |
To get started securely and effectively:
- Update your Twilio credentials in
twilio_creds.py - Place your YOLOv5 models in the
models/folder - Ensure
key.json(Firebase service account) is valid - Add sensitive files and directories to
.gitignoreto avoid accidental commits
*.pyc
__pycache__/
venv/
twilio_creds.py
key.json
models/- Enhanced multi-camera support
- Deploying on cloud infrastructure for continuous uptime
- Further model tuning for diverse weapon types
- Development of a centralized dashboard with log insights and alerts