Improved automatic number plate recognition system

One of the main applications of pattern recognition is the use of video or imaging cameras in order to detect and recognize the vehicle license plate numbers. This is important mainly for access, traffic surveillance and law enforcement. Several systems are implemented and working in situ with more or less precision. These systems are based on the use of a feature set extracted from the best video sequences and further recognized using a classifier. In this paper we present a new system based on multiple features and classifiers, The classification results are further used as inputs for a decision fusion. The classification results are proven to be significantly better that each of the method considered individually. The individual classifiers as well as the decision fusion are tested on a large real image database. Key–Words: Licence plate, ANPR, Pattern recognition, Neural networks, Skeleton features, Matching templates, Decision fusion