Penerapan YOLO dan OCR untuk Deteksi dan Klasifikasi Plat Nomor Kendaraan di Universitas Malikussaleh

Abstract

Automatic vehicle license plate recognition plays a vital role in advancing intelligent transportation systems, especially in enhancing traffic monitoring, parking automation, and security management. This study explores the implementation of the You Only Look Once (YOLO) algorithm and Optical Character Recognition (OCR) for detecting and classifying vehicle license plates based on their regional codes at Malikussaleh University. YOLO is used for real-time detection of license plate regions, while EasyOCR extracts alphanumeric characters from preprocessed plate images (grayscale, sharpening, thresholding). The OCR output undergoes character correction and format validation using regular expressions. The recognized codes are then mapped to regional origins based on official Indonesian license plate regulations. Experimental results show that the system effectively detects and classifies license plates with high accuracy, even under diverse image conditions. Beyond enhancing vehicle identification efficiency, the system offers potential applications in automated campus surveillance and broader public area monitoring. This research contributes to the development of machine learning-based recognition systems and paves the way for future studies in predictive traffic analytics and smart transportation infrastructure.

Keywords: YOLO, OCR, license plate detection, regional classification, vehicle identification