License Plate Recognition for Customer Identification - AutoNation -

Our objective is delivering an improved customer experience by greeting the customer by name, knowing if they have an appointment, being able to review service history, and speeding up the check-in process. We would like to leverage a cloud-based license plate recognition solution and cost-effective cameras that are mounted at the entrance of our service lanes. The cameras will find license plates in the lane and notify the service advisor with all the customers' information.

What it does

The application implements a RESTful API which receives images of vehicles from the client, which then gets processed to get the license plate numbers. Once the software determines the characters on the license plate it runs it across a MongoDB database of dummy customers. For the demo we made it extract simple data like the name of the customer but the same concept can be applied to present employees with data pertaining to the customer such as: appointments, mechanical issues with the car, etc.

How we built it

The backend is being hosted on Google Cloud's Compute Engine, which runs on both Nodejs and Python. We decided that the front-end should be rather robust so we went with a nice and simple Tkinter UI. We used Imgur's API to upload the images to the web to then transfer to the server rather than pushing large images directly to the server. Images are processed super quickly and with extreme accuracy.

Challenges we ran into

This was a first for all three of us on the team so it was definitely a learning experience, we took on technologies and methods we weren't familiar with and succeeded. We ran into several walls here and there, whether it was with MongoDB, Tkinter, Google Cloud, Nodejs, malnutrition, lack of sleep, it was all in good fun.

Accomplishments that we're proud of

We finished. It works great. We are happy.

What we learned

We learned many new technologies and algorithms along the way.

What's next for ShellHacks-Autonation-LPFinder

Better documentation for any other curious people out there that might want to learn more.

We wish to also be considered for the best use of Google cloud, license plate recognition challenge by AutoNation.

Share this project: