Multi-Style License Plate Recognition using Artificial Neural Network for Indian Vehicles

License Plate Recognition (LPR) has gained much focus in the last few decades along with the improvement of digital cameras and the gain in computational capacity. But, there is less work on multi-style license plate recognition (MSLPR). A novel methodology for detection of vehicle license plate with different font styles is proposed in this paper on the basis of adaptive image segmentation in combination with character recognition using Artificial Neural Network (ANN). In the proposed method, plate localization and character segmentation is achieved by using morphological operations in combination with horizontal and vertical edge histogram. Character recognition is achieved by using a two layer feedforward backpropagation ANN. Overall accuracy of 89.5% is achieved by using proposed MSLPR system.