Sensor Tag Detection, Tracking and Recognition for AR Application

Abstract

Augmented reality (AR) applications are widely used in our world today. It may look at the first glance that AR is primarily used in entertainment and learning applications, but in fact, now big companies and corporations try to integrate AR technologies into their industrial processes. It reduces the number of unnecessary communication and increases the net profit of the company. Most companies use AR devices for frontend tasks and a server side for backend. This approach allows reducing the performance requirements for an AR device by implementing time-consuming operations on the powerful server. However, some operations should be implemented directly on the AR device and we have to implement them rationally in terms of performance. Also we have to make interaction between the AR device and a server more effectively. The most important and frequently used operations in AR applications are related to computer vision: image processing, objects detection, tracking and recognition. Therefore, it is important to compare the performance of these operations on AR devices and suggest those that will be implemented on the server side and on the AR device. This paper focuses on deciding these tasks for an AR client-side application for the Siemens company. This application provides a more effective way for monitoring the equipment at remote facilities. Different approaches for sensor tag detection and tracking on a real time video are considered and evaluated. The way for communication between the AR device and the server side is suggested