accident monitoring

Automatic detection and reporting of road accidents using a machine learning and vision algorithm. Built as a Smart City project at HackIIITD.

15
10
JavaScript

VAHAN

Accident Monitoring system for cities.

VAHAN (Vehicle Accident Hazard Notifier) is a web app that notifies the driver whenever there is an accident on the road. Along with that it also sends a message to the nearest police station and hospital with an accurate location of the accident.

  • We aim to deploy our model using cameras on highways in a smart city.
  • Bird’s eye view of the traffic is considered while predicting the collision of vehicles but still our model works well enough on different views as well.

Technologies Used:

  • Tensorflow object Detection API (object detection)
  • Socket.io (for real time communication)
  • Node.js (sever side programming)

How to use :

Clone the repository and then :

  • Crash Detection

    • Setup the object detection API
    • Run the crash_detection_model.py file using python3.
    • Ensure that you have the following libraries pre-installed.
      • numpy
      • six
      • tarfile
      • tensorflow
      • zipfile
      • requests
      • json
      • time
      • PIL
      • matplotlib
      • cv2
  • Server Side

    Run the following commands in the terminal:

    npm install  
    nodemon server/server.js
    

Demo :

DEMO GIF