Skip to content

An end-to-end implementation of a computer vision model for tracking vehicles from video.

Notifications You must be signed in to change notification settings

ymzhu99/traffic_flow_counter

 
 

Repository files navigation

Traffic flow counter 🚦

Introduction

Hello everyone! The following code will be a current work-in-progress app for traffic flow counting. My hope is to make this something of value for my city to be used for traffic flow management.This uses the yolo-v3 computer vision model to vehicles.

Setup

To run the app locally, install the necessary python packages by:

pip install -r requirements.txt

For better reproducibility, make sure to open up a python environment using virtualenv or any of your favorite python environment packages.

The app contains a custom slider. To use it, make sure you have npm installed and then run the following commands.

cd components/custom_slider/frontend/
npm install 

Afterwards, run npm run build

This should recreate the build package for the custom slider.

pull the necessary models and data by running:

make dependencies 

Run the app!

streamlit run app/streamlit-app.py

Future Developments

Here are some of my ideas that would be neat additions to the API:

  • Speed measurement
  • Vehicle distribution measurement
  • Traffic density

About

An end-to-end implementation of a computer vision model for tracking vehicles from video.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 85.2%
  • TypeScript 7.7%
  • HTML 3.2%
  • Dockerfile 2.1%
  • Makefile 1.4%
  • Shell 0.4%