Estimating People Flow Using Point Crowd Density

Abstract

Human location information is highly private, and it is becoming increasingly difficult to obtain human flow information based on an individual's movement history. In this paper, we propose a method for estimating human flows using point crowd density data that can be obtained on a large scale. The proposed method uses an inverse simulation method based on human movement simulation to estimate the movement trend between locations. Through verification experiments using real data, we show that the proposed method can estimate the human flow more accurately than the extended method of previous studies. In addition, through the application to real data and analysis, the practicality of the proposed method is demonstrated.