A smart security system designed for institutions to enhance surveillance and monitoring.
Problem Statement
Ensuring the detection and notification of unauthorized events on school premises using multiple CCTV feeds.
Solution
Intrusion Detection
We utilize a waterfall algorithm to detect human intrusion on the premises.
- Frame Differencing: Detects movement by analyzing differences between consecutive frames.
- Object Detection: When suspicious activity is detected, the YOLO11s model identifies objects in the frame and notifies the authorities accordingly.
Inference Area Selection
Users can define custom boundaries to monitor only specific areas within their institution, reducing unnecessary alerts.
Automatic Scheduler
The system allows users to schedule active hours when surveillance should be disabled, ensuring that monitoring is paused during institution working hours.
Real-time Notifications
Users can customize notification preferences and select specific events for alerts. Respondents receive a secure URL with incident details, enabling them to take immediate action.
Remote Incident Resolution
In the event of an incident, users can use remote links to view snapshot of the incident, communicate with on-site personnel, and resolve the situation.
Key Features
✔️ Human Intrusion Detection
✔️ Customizable Inference Area
✔️ Automated Scheduling
✔️ Real-time Alerts with Remote Access
✔️ Incident Resolution via Remote Interface
Tech Stack
- Django, Python, Twilio, Tailwind, OpenCV, Ultralytics
Upcoming Features
🔹 Additional event detection (e.g., fights, unattended objects)
🔹 Remote system maintenance and monitoring
🔹 Facial recognition for authorized personnel
🔹 Integration with local law enforcement for high-priority incidents