Comparison of different wavelets for automatic identification of vehicle license plate

With increasing number of vehicles in developing countries, the traditional practice of manual monitoring of vehicles is becoming cumbersome, ineffective and economically unviable. This study uses an image-processing-based frequency-domain approach using wavelet multiresolution analysis (MRA) to overcome the difficulties associated with the conventional approach of employing manual observation for vehicle identification. The classification algorithm uses features extracted from an image of vehicle license plate (VLP). Wavelet MRA technique is used to extract the features of image exploiting its abrupt change of intensities. As the features of an image are wavelet dependent, a number of wavelets have been used to extract features of the same image of a VLP. The wavelet that results in features for distinct classification is selected for this application. The case studies pertaining to vehicles in India validate the efficacy of the proposed methodology.