Implementation Of YOLO Through DevOps for Automated Vehicle Number Plate Recognition

The use of automatic number plate recognition (ANPR) systems has gained popularity in recent years including traffic control, law enforcement, toll collection as well as management of parking. However, older ways always have problems with scale, deployment and support. In this paper, we describe how to implement the You Only Look Once (YOLO) object detection model with the help of DevOps principles in order to increase the automation, accuracy and scale of ANPR systems. In this study, the authors relied on the YOLO real-time detection algorithm to implement the core machine learning model for detecting and reading number plates of vehicles with high accuracy. As a result of DevOps practices, the use of YOLO in a DevOps pipeline provides the ability to continuously integrate, develop, test deploy and update the model as new datasets are provided to the system. This approach uses containerizing (Docker) and orchestration (Kubernetes) for scalable application deployment and used CI/CD tools including Jenkins and GitLab CI for automated build and deployment processes. The authors also integrated system monitoring tools on the system to receive timely information, improve performance, and resolve adverse events to ensure the desired level of system reliability and availability. Keywords: YOLO, vehicle number plate recognition, DevOps, real-time applications, continuous integration, continuous delivery, rapid deployable, containerization, IaC, smart cities, intelligent transport systems.