Three-Stage License Plate Detection Based on YOLOv8
Vehicle license plate recognition is of great significance in intelligent transportation systems.To address the issue of false positives in vehicle license plate detection,this paper pro-poses a three-stage method for license plate detection in natural scenes.The proposed method consists of three main steps.Firstly,the YOLOv8 algorithm is used to detect the vehicle's loca-tion and mark it.Secondly,the marked vehicle regions are used to detect regions of interest for text,from which potential license plate samples are obtained and recognized.Finally,a vehicle li-cense plate database is established,and the recognition information obtained in the second step is compared to determine the vehicle's license plate.Experimental results show that the three-stage vehicle license plate detection method proposed in this paper effectively reduces false positive is-sues in license plate detection and achieves good performance in terms of accuracy and speed,with a license plate detection accuracy of 97.5％.This provides an effective approach for subse-quent vehicle license plate recognition tasks.
vehicle license plate recognitionlicense plate detectionYOLOv8regions of interest
万方数据
维普
