Enhanced Automatic License Plate Detection and Recognition Using Yolov11

In low-income nations, traffic accidents brought on by seat belt infractions continue to be a major problem. Finding the cars engaged in these infractions is essential to solving this. In order to increase traffic safety, this paper proposes a method for detecting and recognizing license plates. The three primary parts of the system are character recognition, plates detection, and windshield detection. The technology uses the You only Look Once the eleventh version (YOLOv11) paradigm for windshield detection. Finding the region of Interest (ROI), and using YOLOv11 are steps in the license plate detecting procedure. The most recent version, YOLOv11, is faster and more efficient than its predecessors, The study's dataset was collected from Kaggle and olx. According to evaluation findings, the YOLOv11n variation performed better than the YOLOv11m and YOLOv8m variants in terms of windshield recognition precision and detection speed. The suggested approach achieved flawless precision and recall for license plate detection. Compared to employing YOLOv11n alone, and YOLOv11 produced greater accuracy and average precision throughout the character recognition phase. In particular, the system recognized characters with an accuracy of 0.956 on an average quality of 0.942. This study shows great promise for automatic and real-time detection of cars that are not wearing seat belts on freeways.