Ensemble Haar and MB-LBP features for license plate detection

Abstract

This paper presents a new license plate detection algorithm, in which, the Haar and MB-LBP features are combined and the updated rules of the sample weights are revised. The cascade classifiers are used to detect digitals in the image, Non-Maximum Suppression and the license plate characteristics are applied to locate license plate area accurately. Experimental results show that the proposed method could effectively avoid the phenomenon of weights distortions and get higher detection rate while reducing false alarm rate. © Springer-Verlag 2013.