Estimation method of the number of pedestrians from surveillance video footage

Abstract

In ubiquitous network society, the information of mobility nodes such as pedestrians with cell phones is very useful for network configuration, communication control and so on. In this paper, we propose an approach to estimate the number of pedestrians using surveillance video footage from a stationary camera. The proposed approach extracts variation regions from video snapshots by using background difference and frame difference, and estimates the number of pedestrians from the area of the extracted region, using conversion algorithm constructed from the statistical data. We evaluate the efficacy of the proposed approach by comparison the estimated values with the true values and find that the mean difference in the number of pedestrians is within 0.36.