Mean shift for accurate number plate detection

This paper presents a robust method for number plate detection, where mean shift segmentation is used to segment color vehicle images into candidate regions. Three features are extracted in order to decide whether a candidate region contains a number plate, namely, rectangularity, aspect ratio, and edge density. Then, the Mahalanobis classifier is used with respect to the above three features to detect number plate regions accurately. The experimental results show that our algorithm produces high robustness and accuracy.