Classification and Recognition of Arabic License Plates

摘要

Extensive road transportation network has increased the necessity of strict rules, security and safety. Intense number of vehicles are difficult to manually deal with. License Plate is the peculiar identification of any vehicle. Recognizing it can be beneficial in various ways, notable of which are Automatic toll collection, parking, criminal tracking, access control, etc. Presented here is the method for classification and recognition of License Plates of Dubai and Saudi Arabia. Each of these countries have some differences and similarities in their plates. Considering the features, the plates are classified according to their countries. The further recognition of letters in it is done by training the Neural Network. Two layers of Feed forward neural network forms a network which uses scaled conjugate gradient Back propagation algorithm for training of network. Experimentation is carried on available standard license plates for these two countries. The new approach is applied for classification and recognition of number plate of two countries, but use of this approach can be expanded for number plates of more countries.