Design flow of vehicle License Plate reader based on RGB color extractor

Abstract

In the vehicle license plate reader system, the License Plate Recognition (LPR) is designed based on machine vision technology without any direct human intervention. This paper presents an efficient LPR technique based on the RGB color extractor. The proposed method is capable of recognizing alphabets and numeric characters on the license plate in real-time. This technique has been tested with a large number of images in order to analyze its performance. The tested images are captured from the front and rear of the vehicles under different conditions, including different angles, luminance, and weather conditions. With the real-time test results shown in this paper, we obtain 98.5% accuracy for character extraction and 95.1% accuracy for character recognition using this technique.