<![CDATA[Manual collection of vehicle license plates is often inefficient and prone to errors, so an automatic identification system is needed. This research aims to implement and evaluate the performance of a license plate character detection system, focusing on the accuracy comparison between black and white base plates in Indonesia. The method used is Optical Character Recognition (OCR) with image preprocessing workflow including Grayscale, Gaussian Blur, and edge detection implemented in Google Colab. The system was tested using 100 primary data samples consisting of 50 black base plates and 50 white base plates. The findings showed that the system achieved a combined average accuracy of 84.36%. Specifically, it was found that the accuracy on the black base plate (85.40%) was slightly superior to that on the white base plate (83.32%). The implication of this study is that the change in license plate standards has a measurable technical impact on the ANPR system, where the findings can serve as a foundation for developers to calibrate the system to be reliable on both plate types during the transition period.]]>
Copyrights © 2025
