A vision-based method to estimate passenger flow in bus

Abstract

In this paper, we describe a procedure using image processing for real-time estimation of passenger crowd flow in a complicated background in a bus. The contour feature of the head is exploited to locate every passenger's position. In order to obtain the contour feature accurately, the modification of the Hough Transform and the fuzzy measure are applied. The procedure has been implemented in a DSP-based embedded system for field tests and the results show that the procedure mentioned here is effective.