Automatic license plate detection based on corner point and cluster

Abstract

This paper introduces an automatic method for licenses plate detection that using local corners points features, clustering and some properties of license plate. Licenses plates are rich corner point's area that can be used with some properties of license plate to locate licenses plate location. The algorithm has four stages; firstly, the quality of image is improved via preprocessing operations. Consequently, Harries corner point's detector is used. These corner points are gathered in clusters, where clustering of corner points were built on relative difference distance that guarantee the whole license plate will be in only one cluster.