A Rapid Crowd Density Estimation based on Proper Reference Sub-Region

Ning Li and Yan Zhang

Keywords

crowd density estimation, image sequence, color background generation, reference sub-region, proportional calibration

Abstract

Color background generation and proportional calibration are present to enhance the developed method of crowd density estimation using a reference sub-region in the paper. A sequence of color crowd images is applied to generate a color background image with the composition of R, G and B components, relatively. An adaptive proportional calibration approach involving calibration of the crowd image as opposed to calibration of the camera is proposed to estimate crowd density at various sites with an established relationship between image parameters and crowd density. According to the relationship between a camera system and an image plane, a special proportion is got to segment the investigated scene into several proper sub-regions. With the established function, the estimation in the reference sub-region can be applied for crowd density estimation in other sub-regions rapidly. Experimental result shows that the proposed method is simple, effective and fast in practical application. The method developed may be used in other surveillance systems.

Important Links:



Go Back