RowdyPlates

Inspiration

If the government can do it how hard can it be

What it does

It has the ability to read and save information from license plates in real time and can pull meaningful correlations from the data

How we built it

Through hosting on a rapspberry pi and extensive python and nodejs libraries

Challenges we ran into

finding a good machine learning platform that accurately and reliably pulls data from images but thanks to my man Matt Hill who apparently single handedly developed and maintained openalpr; the best license plate recognition software in the game

Accomplishments that we're proud of

reaching the goals that we set out to accomplish

What we learned

never use nodejs, if an api is over 3 years old just don't do it even if it looks so good, dont subprocess what you can thread, also don't sub process never ending loops

What's next for RowdyPlates

hopefully getting eyed by some three letter agencies

Built With

Submitted to

RowdyHacks 2019

Created by

I helped develop the main Python script that ran the application, aided in the integration of the machine learning model, helped with general debugging and wherever else I was needed.

Shane Barnett
I worked on the licensePlateVehicleInfo.py which included scraping a HTML document to parse data. This was my first time experiencing web scraping and after research I found BeautifulSoup to be an efficient way to do so. The .py also made directories for each license plate number. Inside each directory includes vehicleHistory.txt (history of license plate captures) and a vehicleInfo.txt (scraped from the HTML) which gives the make, model, year, body, and engine. Overall I felt we pushed through many barriers as a team and as an individual I couldn't have had a better time learning.

bairdChris
Robert Gilliam
nickbrell