Video analysis of vehicular flows for road traffic monitoring

• Main Author: Ong, Kai Sin.
• Other Authors: Ang Yew Hock
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision; DRNTU::Engineering::Civil engineering::Transportation
• Online Access: http://hdl.handle.net/10356/14741
• Institution: Nanyang Technological University

Video Analysis of Vehicular Flows for Road Traffic Monitoring” is a final year project that aims to develop a real time computer vision system that can collect traffic data, detect traffic incidents and has practical applications on intelligent traffic monitoring. Videos are sourced from Land Transport Authority (LTA) and the objectives of this project is to create a program that can detect traffic events through streaming from the IP network camera and alert the end user on the traffic conditions on the particular highway. In this report, theoretical and technical aspects of computer vision techniques are explained in detail. In theoretical aspects of computer vision techniques, techniques that will be explained in depth will be grayscaling, image differencing, edge detection and automatic thresholding. Some techniques have been implemented in the proposed system. Also, technical aspects of the prototype will be discussed. The prototype is able to retrieve images from the live streaming via the network camera and preprocess the images using image differencing. Foreground extraction is retrieved and pixel intensity is calculated by the system which serves as a guide and signal for traffic flow at the point of time. This helps to classify the congestion level.