Application of Image Processing to the Analysis of Congested Traffic

Images are one of the most common and convenient means of sharing information as they provide accurate data about positions, sizes and relationships between objects. Due to this, many image processing applications can be found... [ view full abstract ]

Images are one of the most common and convenient means of sharing information as they provide accurate data about positions, sizes and relationships between objects. Due to this, many image processing applications can be found in the literature in different fields, such as taxonomy or microscope image processing in Biology, forensics, cryptography and biometrics in Security or traffic monitoring and metallurgy in Engineering.
The aim of this paper is to explore the capabilities of using images for applications on transport, especially for traffic monitoring. In that sense, a high resolution camera will be used in this work in order to capture aerial images of congested traffic. These images will be processed to distinguish all vehicles as different objects on the road and to identify the type of vehicles (regular cars or trucks). In order to achieve that result, an algorithm able to detect points at each vehicle’s edges will be firstly applied. Each of these sets of points will be then classified into different clusters, each of which corresponds to a different vehicle on the road.
With all of the information from this image processing procedure, a model of the distribution of traffic will be obtained, and so the distribution of loads on the road can be analysed.