Advanced Surveillance System

Project Overview

This project was developed as part of the CodeCrafter Hackathon, where the challenge was to build an advanced surveillance system using AI/ML and various sensors like cameras, lasers, radar, and fiber optics. Initially stunned by the complexity of the task, our team transformed the challenge into an opportunity, creating a high-security headquarters simulation on a dynamic map that adapts in real-time to environmental conditions. The journey from hesitant beginnings to a fully functional system, culminating in a 3rd runner-up podium finish, was intense but incredibly rewarding.

Team

• Soham Bhaye
• Pushkar Mhatre
• Sujal Shah
• Oshnikdeep Tiwari

Features

1. Object Detection

• Our system detects objects only when the laser is interrupted, ensuring precision and reducing false alarms.
• Image Demo:
  (Shows a soldier crossing a laser grid with detection events logged.)

2. Facial Authentication

• Only authenticated users with a unique facial identity can enter the headquarters.
• Image Demo:
  
  (Displays a standard camera feed with a "Verdict: Authorized" result.)

3. Environmental Adaptation

• The system adapts based on weather and visibility conditions (e.g., rain, fog, day, and night) to determine the best-performing sensor in real-time.
• Image Demo:
  
  (Displays a thermal camera feed with device controls for different conditions.)

4. Sensor Fusion

• Integrated data from multiple sensors (camera, laser, radar, and fiber optic) to enhance accuracy and decision-making.
• Image Demo:
  
  (Shows a dynamic map with overlapping sensor ranges in green, blue, and red.)

5. Weapon Detection

• The system identifies potential weapons (e.g., knives) in real-time using AI-based object recognition, triggering a "BREACH DETECTED" alert when detected.
• Image Demo:
  
  (Shows a person holding a knife with a breach alert and accuracy score.)

6. Buried Fiber Optic Frequency

• Utilizes buried fiber optic sensors to detect vibrations or movements, providing frequency data to identify potential intrusions or activities underground.
• Image Demo:
  
  (Displays a graph of fiber optic frequency data overlaid on a map.)

7. Laser Simulation

• Tracks objects crossing the laser barrier, calculating distance and speed to assess potential threats.
• Image Demo:
  
  (Shows a laser grid with a soldier and speed estimation details.)

8. Speech Recognition for Alerts

• Implements real-time speech recognition to generate alerts based on specific threat-related keywords or commands, ensuring quick response times.
• Note: This feature is demonstrated through logged events and can be tested with audio input in the system.

9. Multilingual Chatbot

• An AI-powered chatbot capable of understanding and responding to queries in multiple languages, making the system more user-friendly and accessible.
• Note: Demo available upon running the chatbot module.

Technologies Used

• Machine Learning: TensorFlow, OpenCV
• Backend: Python (Flask/FastAPI)
• Frontend: React.js, TypeScript, Tailwind CSS, ShadCN
• Database: PostgreSQL
• Authentication: Clerk (3FA Authentication System)