Multi-Camera Video Surveillance

Abstract

In this paper we aim to detect and track moving persons inside a target area and to obtain statistics regarding the regions of the area with the highest pedestrian traffic. We present or approach on video data gathering, storage and analysis, aiming to deliver real-time results. We use neural networks for both pedestrian detection and pedestrian tracking, while attempting to track persons between multiple cameras as well. We present and comment our experimental results, pointing out advantages and disadvantages.