Certain Investigations on Vehicles Number Plate Identification using Top Hat Transform Algorithm and Optical Character Recognition

Abstract

Investigation on vehicles number plate with top hat transforms is the method to recognize the characters on number plate utilizing the process like Image processing and OCR. The conception of this project is, first the image of the vehicle is to be captured. Next, the number plate of the vehicles is extracted from captured images using Top Hat transform algorithms. Conclusively, Optical Character Recognization recognizes the character presented in number plate. Additionally, the extracted data is stored in our database. This project can be implemented on various security zones like Parking Systems, Traffic Control areas, Toll gates, tracking of vehicles, etc. In the current scenario, the usage of vehicles increases day by day. Hence it's impossible to maintain the record manually for entire Vehicles. By expanding this system it becomes easy to sustain such rather records. In the majority of the nations, the extent of the number plate relies upon the aspect ratio. It can be figured by Width over Height. This work proposes the strategy for following Indian Number Plates of the vehicle. While contrasting other number plate extraction strategy this technique varies in such a path, in several strategies, they utilized just an area of a number plate for recognizing the character. However, in this method the entire vehicle can be included which first finds the particular zone of number plate then it executes character Recognition. Template matching technique where used in previous methods of number plate identification which one and only needs an area of a number plate. The disadvantages of previous techniques are it can only recognize already stored character in the templates and if there is more than one number plate, it is impossible to identify the sector of the number plate. Therefore to overcome such errors, we developed this algorithm which relatively gives better results while comparing with other methods. The absolute time taken for one execution is below 5 seconds.