Indoor Tracking for Large Area Industrial Mixed Reality

Abstract

For mixed reality (MR) applications the tracking of a video camera in a rapidly changing large environment with several hundred square meters still represents a challenging task. In contrast to an installation in a laboratory, industrial scenarios like a running factory, require minimal setup, calibration or training times of a tracking system and merely minimal changes of the environment. This paper presents a tracking system to compute the pose of a video camera mounted on a mobile carriage like device in very large indoor environments, consisting of several hundred square meters. The carriage is equipped with a touch sensitive monitor to display a live augmentation. The tracking system is based on an infrared laser device, that detects at least three out of a few retroreflective targets in the environment and compares actual target measurements with a precalibrated 2D target map. The device passes a 2D position and orientation. To obtain a six degree of freedom (DOF) pose a coordinate system adjustment method is presented, that determines the transformation between the 2D laser tracker and the image sensor of a camera. To analyse the different error sources leading to the overall error the accuracy of the system is evaluated in a controlled laboratory setup. Beyond that, an evaluation of the system in a large factory building is shown, as well as the application of the system for industrial MR discrepancy checks of complete factory buildings. Finally, the utility of the 2D scanning capabilities of the laser in conjuction with a virtually generated 2D map of the 3D model of a factory is demonstrated for MR discrepancy checks.