A system design for detecting moving objects in capturing video images using laser range scanners

Abstract

When video images are obtained from a camera in an IoT environment, preliminary processing, such as the addition of metadata, may be conducted for smoother and more interactive data usage. However, it is often difficult to perform high-load processing, such as video image analysis to obtain metadata, in the case of general camera devices, which typically have lower processing capability than personal computers. While it is possible to add metadata by using an edge device that can conduct desired calculations on a communication pathway (such as a router with edge computation functionality connected to a camera), this approach may not be sufficient for high-load processing. Against this backdrop, we propose a technique that enables preliminary processing on an edge device using a variety of external sensor information, rather than conducting high-load processing. As a case study, we discuss a video system that employs a laser range scanner for motion detection and an edge device that coordinates with a camera to add the positions of moving objects as metadata on a video image.