College Bus Number Plate Registration Detection Using YOLO-V8

There is a need for automation of vehicle entering system and to reduce the time of verification. We propose a solution for this problem considering college buses o BVRIT as organization. College bus number plate Registration Detection is crucial part of smart BVRIT planning and BVRIT transport management. This paper presents YOLO V8 algorithm for Detecting college bus using number plate registration detection. YOLO - You Only Look Once. This YOLO-V8 algorithm is state-of-the-art (SOA) model or algorithm. It's real-time object detection algorithm that can detect objects in any image, that has many classes in it with high accuracy and speed. A dataset of College bus images is collected manually and finetuned using transfer learning. Model is trained on powerful GPU (Graphics Processing Unit), significantly reducing training time and taking less time for Character recognition. The model's accuracy is computed by average, precision, recall, and F1 score as well curves. Our results show that Yolo V8, model can accurately detect the bus with high accuracy and speed in any weather condition. The text from detected number plate is recognized using Easy-OCR. This number plate text is used to infer whether bus is of COLLEGE BUS or NON-COLLEGE BUS. Model's overall performance is remarkable. While maintaining recall of 95 percent, which improves its capacity to collect a substantial number of pertinent things, it attained a precision of 99.2 percent, suggesting a high accuracy in properly identifying objects. The model's ability to detect items with a high degree of confidence is demonstrated by its mAP50 score of 99 Percent. This is one of the optimal solutions to minimize human work and time, making automation. This research has been done with the purpose to find the registered college bus is supposed to enter the college campus and Non-college bus is not supposed to enter the campus.