Automatic Number Plate Recognition Using Convolution Neural Network

Automatic number plate recognition system (ANPR) finds the vehicle that violates traffic signal by detecting from the number plate of the vehicle. After finding the details of the vehicle, it sends a violation message to the owner. This ANPR will take the digital image as input. It uses segmentation techniques to detect the number plate region. The segmented image is compared to the original number plate image by correlation technique. The obtained data is checked with the database. We used MATLAB software which extracts the number plate from the input image, and it further uses GSM module (SIM900) to send a violation message to the person after violating traffic signal. Here the input image is further processed using convolution neural network algorithm for more accuracy purpose. After segmentation, each character made to compare with the thousands of characters for more accuracy. After obtaining the number plate number, it checks the data base for details of the owner. Later the process sends a violation message to the owner of the vehicle. Foreign ANPR systems are not able to connect with the Indian systems due to less compatibility. So this paper decreases the difficulty of the system.