Pengesan lalu lintas dengan kaedah pemerosesan imej

Abstract

Most traffic detectors in service today work based on inductive loops buried under the road surface. Recently cameras and image processing with high-speed computers were beginning to be adopted for traffic surveillance system. This study was conducted to develop efficient traffic detector using image-processing with a personal compute I: Traffic images were captured using ordinary analogue video camera and digitised using a video blastel: Four algorithms were developed i. e. algorithm for traffic counting, algorithm for vehicle classification. algorithm to determine the speed of the counted vehicles and algorithm for incident detection. The four algorithms were found to be sufficiently efficient to be executed by a personal computer to capture multiple road lane traffic data simultaneously with high accuracy both during daytime and during, night