Inspiration

Truck drivers drive for long hours, and are often very sleep-deprived. This can lead to accidents, and a lot of them. Approximately 500,000 accidents are caused by trucks: this is a serious issue that needs to be fixed. We wanted to create a system that alerts the driver when they are sleepy as well as points directions to the nearest rest stop using a training model which measures if they are sleepy or not judging from their eyes and their yawn, getting an image from their webcam. Then, these accidents could be easily avoided.

What it does

A model deems a person as sleepy judging their eyes and their yawn from the webcam. Then, the code for the nearest location(obtained using geocode API) for a rest stop is printed to the site as well as directions(google directions API). A interactive map is also incorporated to make it easier for instructions to a rest stop nearby.

How we built it

We used Flask for the backend and to integrate onto the website, TensorFlow, openCV, dlib for the model, and different google APIs(i.e. geocoding, directions, maps, google maps) to incorporate the location and the map into our sight.

What's next for SafeDrive

  1. Deploy this app
  2. Clean up presentation of app make it more ideal for user
Share this project:

Updates