License Plate Reader

Inspiration

The inspiration for this project was motivated from the growing need for efficient and automated vehicle identification systems whether in parking management and toll collection and, most importantly, law enforcement and crime prevention. By utilizing machine learning models, we are aiming to create a solution that will enhance reduce human error, further streamline and optimize operations, and, most importantly, enhance overall security.

What it does

Our project develops an automated license plate detection system in which machine learning is utilized to identify and extract license plates whether through images or video feeds. The system we have created is designed to work in real-time to provide quick and reliable recognition.

Challenges we ran into

Finetuning the machine learning model (errors in consistency, accuracy, etc. in test results)
Understanding AWS architecture such as AWS Rekognition, Sagemaker, E3 Instances, and Buckets

Accomplishments that we're proud of

Navigating both the AWS architecture as well as documentation regarding third-party libraries that we have used to create this system such as OpenCV and EasyOCR

What we learned

How to train an ML model
How to identify parameters within an ML model

What's next for License Plate Reader

Hopefully as we continue to develop and improve upon this project, not only will we be able to detect and read license plates but, additionally, we will be able to detect and read a specific/targeted license plate in various images in which we can capture through a bounding box.