Abstract
License plate localization is considered as the most important part of license
plate recognition system. The high accuracy rate of license plate recognition is depended on
the ability of license plate detection. This paper presents a novel method for license plate
localization bases on license plate features. This proposed method consists of two main
processes. First, candidate regions extraction step, Sobel operator is applied to obtain
vertical edges and then potential candidate regions are extracted by deploying mathematical
morphology operations 5. Last, license plate verification step, this step employs the
standard deviation of license plate features to confirm license plate position. The
experimental results show that the proposed method can achieve high quality license plate
localization results with high accuracy rate of 98.26 %.
Users
Please
log in to take part in the discussion (add own reviews or comments).