Vehicle Identification Systems using Virtual Line Sensors and Speed Up Robust Features

Abstract

This work proposes an improved and robust algorithm to a virtual line approach object movement detection using computer vision technique. The improved performance will be emphasized on how one line could detect multiple object, determines their sizes and could at most classify the object into group similarities. The algorithm later will identify the vehicles based on their license plate using speed up robust features - SURF technique. The method incorporating Bresenham’s points plotter algorithm to determine the points coordinate in a line array, which later calculating movement within the given points and group neighboring points with pixel values indicating movement that are in close proximity as one object. This could then calculate the number of points laid within the group which then could determine the size of the object. As a result of the proposed work, a robust object classification technique that is implemented as a computer application that can be used in a live environment context.