CO2View

See a car's carbon dioxide emissions, carbon monoxide emissions, and more.

Inspiration

Over 75% of carbon monoxide pollution, and 27% of greenhouse gas emissions are caused by cars. We were inspired to apply computer vision to help improve peoples' understanding of vehicle emissions, and to gather data that may provide insight into what areas are most negatively impacted.

What it does

Users submit an image of a car, or enter in their plate number, to see details about its emissions, and how it compares 13,000 other unique car models.

How we built it

CO2View.com runs on Google Cloud, using Django as a backend framework and a React.js frontend. We created Django REST API's for processing submitted images and for retrieving carbon emissions data from license plate numbers. We used an API for retrieving the number and state from zoomed-in images of license plates. We wrote all image preparation and transformation using OpenCV.

What's next for CO2View

Our website currently uses our API's to provide individuals with vehicle emissions information. However, these API's could also be used to gather data on emissions in a certain location, through the use of a Raspberry Pi with a camera and the YOLOv3 object detection model. While we demonstrate a proof-of-concept of this second use, we believe it could be greatly expanded to understand more about the sources of air pollution in the future.