A metrological approach for Augmented Reality tooltip tracking assessment
提出一种定量计量学方法，用于评估增强现实跟踪系统的工具提示跟踪性能，通过3D打印测量工件和不确定度分析，比较了HoloLens 2和立体相机系统的精度。
Tracking systems are essential in various fields, such as health and manufacturing industries, enabling mapping between the real and digital worlds. Amongst others, Augmented Reality Tracking Systems (ARTS) are more recent and less explored. This work proposes a quantitative metrological methodology to evaluate ARTS tooltip tracking performance, facilitating benchmarking, parameter optimization, and system selection for specific tasks. A specific 3D-printed measuring artifact is proposed to guide tooltip positioning. Tracking accuracy and precision are estimated, highlighting the effects of influence factors. The methodology was tested with two commercial state-of-the-art ARTSs using marker-based tooltips, i.e., a Microsoft HoloLens 2 and a stereo camera system equipped with Intel RealSense SR305 cameras. Metrological characteristics are evaluated, and the Euclidean distance expanded uncertainty at a conventional 95% confidence level is estimated as 5 . 071 mm for the HoloLens 2 and 6 . 800 mm for the stereo system, resulting in a superior metrological performance of HoloLens 2 under the specified conditions. This study provides a standardized approach for quantitatively comparing AR tracking systems, offering valuable insights for optimizing their use in specific applications and, innovatively in the context of ARTS, associates measurement uncertainty with tracked distance values.