Abstract

The rapid urbanization has witnessed exponential growth in the vehicular movement, becoming difficult for the authorities to verify insurance details, traffic violators and tracking stolen vehicles on the road. The proposed work captures the vehicle registration plate image using a standard resolution camera for detection and recognition using a Deep Neural Network text detector model. The feature extraction of this framework is enhanced by a faster R-CNN version for scene textual data called PVANet. Automating the detection, recognition and notification of miscreant vehicles could ease the cumbersome process of tracking vehicles in signals, toll booths, private and government restricted areas. The registration number extracted is then crossreferenced using RPA, with the data provided by the insurance company and police department to identify defaulters with automatic notification to the concerned authorities without any human intervention