Ice on bridges lead to dangerous traffic conditions. This is continued work on the icy bridge project where live camera images from NCDOT are analyzed using image processing and machine learning.

What it does

Examines cameras imagery, analyzes and alerts when ice is detected.

How we built it

Microservices hosted by AWS using Javascript, python, bash and jupyter/sagemaker notebooks.

Challenges we ran into

  • Changing camera imagery is always problematic.

  • Lack of participants.

  • Once, I found out the instances did not support mxnet, I knew I would not be in the bounds. I punted on the project in terms of this hackathon. I'm still working on the mxnet portion though.

Accomplishments that we're proud of


What we learned


What's next for Icy Bridge Identification


Built With

Share this project: