Project Overview
Focused on accurately identifying and recording vehicle license plates, this initiative aims to streamline entry and exit processes within parking lots. Utilizing advanced image processing and pattern recognition algorithms, the system can efficiently recognize license plates, facilitating seamless access control and improving overall security. The system is combined of multiple deep learning models, which includes YOLOv5, WPOD-Net and OCR.
The system is deployed on NVIDIA Jetson Xavier device and being applied efficiently by our clients in both Korean and Russian.
Responsibilities
Experimented, error analysed and intergrated multiple models (YOLOv5, WPOD-Net, ...), resulted in more than 98% in term of accuracy.
Deployed trained model on NVIDIA Jetson Xavier device. Built server-client protocol to verified system pipeline and dockerized product.