Vehicle Number Validation System Using Convolution Neural Networks

Abstract

Recording and maintaining registration number of vehicles at public parking lots, and organizations, hospitals and educational institutions are important for the security of the organization to keep a track of incoming and outgoing vehicles. Manual entry of the registration numbers are tedious, generally not accurate and moreover the records are not tamper proof, which raises serious concerns about security and accuracy.The proposed system aims to solve the above problem by using an optical character recognition system based on a convolution neural network that is trained to interpret characters present on a vehicle’s license plate. The system can be programmed to allow only vehicles that have their license numbers registered with the organization. The system is programmed to instantly send out SMS to the registered user’s mobiles number upon successful access into and exit out of the organization. This enhances the security further and helps in preventing unauthorized access into the organization.