<![CDATA[This study develops an automatic system for vehicle license plate identification in Indonesia by combining the YOLOv8 object detection algorithm with character recognition using Tesseract OCR. The system is designed to recognize two commonly used license plate colors in Indonesia: white and black. The process begins with training the YOLOv8 model using an annotated dataset consisting of 602 images of Indonesian vehicle license plates. Once the license plate area is accurately detected, the image is cropped based on the bounding box coordinates for the character recognition stage. Before being recognized by Tesseract OCR, the cropped image undergoes a series of preprocessing steps to enhance recognition accuracy. These steps include grayscale conversion, local contrast enhancement using the CLAHE method, segmentation with Otsu thresholding, and color inversion using the bitwise NOT operation. The system is evaluated using metrics such as accuracy, precision, recall, and F1-score. Test results show that the system is capable of detecting and recognizing license plates with high accuracy achieving 96% for white plates and 92% for black plates. Overall, the system achieved a combined accuracy of 94%. These findings indicate that the integration of YOLOv8 and Tesseract OCR, supported by effective image preprocessing, can significantly enhance license plate recognition performance. Nonetheless, challenges remain in recognizing characters on black plates with low contrast, which could be the focus of future improvements.]]>
Copyrights © 2026
