A hybrid machine learning based approach for Hindi license plate recognition

Abstract

In the present era of revolutionary techniques, the developed countries are using various technologies to provide the better quality of social life. In these technologies the automatic number plate recognition is playing vital role in intelligent transportation management. Most state-of-the-art ANPR are compatible with language specific license plates specifically English. The langue-specific characteristics limit the competence of ANPR where more than one language is used in license plates. In this paper a system for recognition of Hindi license plates is proposed. The proposed approach is based on hybrid machine learning techniques such as support vector machine and artificial neural network. These two different machine learning approachs are integrated in order to improve the performance of system. The average license plate recognition accuracy of proposed approach is above 96% in recognition of Hindi license plates.