A novel image segmentation technology in intelligent traffic light control systems

Abstract

This paper analyzes the present application of intelligent traffic light control systems and points out the advantages of controlling traffic lights according to the traffic information collected using video cameras. A novel method that recognizes the detection area based on Hough transformation, and computes the number of vehicles based on the detected boundaries between the windshields and the hoods is proposed. Furthermore, the time complexity of the image erosion algorithm involved is improved. Experiments indicate that the proposed procedure can achieve fast and high-precision recognition for crowded vehicles thus is efficient in practice.