Vehicle Detection System.

Abstract

VDS, is a real-time system that is used to detect vehicles and their license plate. It uses the live feed from the camera and divides them into individual frames. These frames are then passed to the two core algorithms of our system that are vehicle detection and license plate detection. VDS takes the frames and passes them to a machine learning model. This model was made using a CNN which then predicts the class of a vehicle present in the frame. Vehicle data like make, model is predicted using the model. The second part of the system, license plate detection, takes the frames and applies image processing techniques like Gray-scale conversion, Normalization, and edge detection. Character Segmentation is also applied to the image. Finally, Optical character recognition is applied to previously extracted segments. Then the license plate number is stored in string form for further use. In the end, the data from the vehicle detection algorithm and license plate detection is displayed to the user.