A New Technique Based on Intensity Distribution Graphs for the Extraction of Traffic Flow Rate and Vehicle Speed

Abstract

This paper introduces a new visual system which allows analysis of traffic behavior on freeways and highways, using a video camera mounted on a relatively high place (such as a bridge). A new technique based on spatial and temporal intensity distribution is introduced to determine the road traffic parameters: average speed (AVSP) and average flow rate (AVFL). After performing many steps including automatic detection of the Region Of Interest (ROI) and camera orientation, background subtraction and Inverse Perspective Mapping (IPM), an Intensity Distribution Graph (IDG) is built for the current frame. The (IDG) is then transformed into a single Grayscale Image Line (GIL). A certain number (N) of GILs representing N frames are concatenated to build a Vehicle Tracking Diagram image (VTD) which gives a whole view of the traffic flow in the ROI during a specific period T. Using the VTD we can visualize the vehicles trajectories, speeds and estimate AVFL. By applying edge detection and Hough Line transform on the VTD, AVSP is estimated. The system has been implemented using MATLAB as an offline application. The results were encouraging, reaching 97% accuracy for the AVSP and AVFL. Further code optimization is needed for online monitoring.