HEAVY TRAFFIC MONITORING IN REAL-TIME
M. Kilger · 1993
This paper presents a video-based system for traffic monitoring. The objective of the system is to detect, track and classify each vehicle even in heavy traffic in real-time on low-cost hardware. The algorithms for the detection of motion, tracking and classification of objects are presented. A novel algorithm for motion detection in heavy traffic is developed, and this is employed when the system detects the need for it. The efficiency of this novel approach is demonstrated by extensive tests on real-time hardware.