In our previous experience with machine learning, we found that a major challenge in the field is the difficulty of finding readily available data. Using the IoT technology, we want to build a data marketplace where various types of data can be traded for personal or business use.

What it does

Databox processes real-time image data from traffic intersections and delivers information on the amount of traffic between any 2 intersections, as well as information on the traffic type, including pedestrians, bike, cars, buses.

How I built it

Databox uses a raspberry pi and the Telus Shield + Tag. The software was developed using YOLO on Pytorch.

Challenges I ran into

We ran into some issues installing Pytorch on raspberry pi and configuring the Telus LTE.

Accomplishments that I'm proud of

We learned the basics of using IoT technology.

What I learned

We learned to use the YOLO model for computer vision.

What's next for Databox

We hope to write an app that processes the data we collected to build a map with information on the traffic between intersections.

Built With

Share this project: