Comet Parking

Inspiration

Finding parking spaces can be a pain during peak hours at the university, we wanted to build something to make that search easier.

What it does

Uses computer vision techniques to detect open parking spaces from images of the lots and classifies a spot as either occupied or free. Images of the lots can be obtained by hooking up a Raspberry Pi with a camera at a high reference point, such as upon a lamppost. For the demo, the images were obtained from Google Maps.

How we built it

We used OpenCV to process the images of the lots and compare every individual parking space with a reference image that represents an empty lot.

Challenges we ran into

Practical difficulties such as working with varying lighting conditions, shadows, different shades of tarmac/concrete in the images obtained.

Accomplishments that we're proud of

Development of an algorithm to detect empty parking spaces with a high degree of accuracy(> 95%).

What we learned

Practical image processing techniques and computer vision using OpenCV.

What's next for Comet Parking

Taking this forward to the Parking and Transportation department of UT Dallas in order to obtain the necessary permissions and funding in order to implement the system at the campus. Building of mobile and web apps to serve the UTD populace with the information obtained from the system.