A License Plate Recognition Method Based on YOLOv8-Pose and E-LPRNet

This paper proposes a license plate recognition method based on YOLOv8-Pose and E-LPRNet for complex scenarios such as urban roadside and road inspection. E-LPRNet is a character recognition network formed by modifying the activation function of the backbone network in LPRNet (License Plate Recognition Network). This method also includes a license plate correction feature, which accurately recognizes license plate characters even when the plates are at large angles, thereby increasing the recognition accuracy. A series of experiments using a large license plate dataset verified the effectiveness of this method.