Moving human head detection for automatic passenger counting system

Abstract

Moving human head detection is the basis of automatic passenger counting system and the speed and the detection accuracy of existing algorithms need to be improved. This paper puts forward a fast and effective algorithm for detecting moving human head. At first, the background is updated by block symmetric difference, then, the moving objects are extracted by culminating symmetric difference and background subtraction, finally, the human head contours are detected using the random Hough transform based on gradient. Experimental results show that this algorithm can detect moving human heads against illumination level changes and extraneous motion efficiently in automatic passenger counting system. © 2012 Springer-Verlag GmbH.