Detecting Crowd through Phone

Abstract

Crowd density monitoring is crucial to many applications, such as guiding tour and crowd control in commercial environments. Video or image-based solutions are high cost and cannot be applied in low-light environments. For some Radio Frequency (RF) based technologies, some people have to carry certain wireless transceivers; others also need to collect abundant finger print. These approaches are high cost and impractical. In this paper, we propose a crowd monitoring approach using mobile phone. Our design of crowd detection adopts clustering methods. Feature sets derive from Wi-Fi signal strength measurements. We use Bluetooth readings analyzing to estimate crowd density. We implement our design on off-the- shelf smart phones and evaluate its performance via extensive experiments in typical real-world scenes. Results of experiment verify the feasibility and the effectiveness of our proposed approach.