License Plate Image Binarization Based on Chromatic Information and Fuzzy C-Mean Approach

Abstract

The license plate recognition system plays an important role in a variety of applications under the umbrella of intelligent transportation systems. However, in the complex real-world road environment, mud, fog, poor illumination, and other environmental situations degrade the quality of the vehicle license plate image and present a challenge to the development of a functional license plate recognition system. To target this problem, a method based on fuzzy binarization is presented. In this approach, chromatic information is applied first to classify each license plate image into four categories that are typical for vehicles in China; then a binarization approach based on the fuzzy c-mean is applied to generate the binary image, during which a dynamic procedure is utilized for adapting the threshold value to the changing real-world road environment. Empirical results based on the processing of 1,000 real-world vehicle license plate images show that the proposed approach performs well compared with traditional approaches.