Page - 108 - in Proceedings of the OAGM&ARW Joint Workshop - Vision, Automation and Robotics

backside of the tablet or an additional AR marker is put on the desired position, see Figure 2. The additional marker Fig. 2. Screenshot of position input via an additional marker in the image. The position of the marker defines the position of the new sticky note. The pane on the right shows pre-defined tags connected with the marker. can be removed after the system took over its position. While the measuring tip might be the more intuitive way to define a position, AR markers have an advantage: they can be combined with note-templates, e.g. different markers for different users or different types of notes. Fig. 3. Screenshot with two sticky notes in the image (flash icons on the left). The top left note is selected and the pane on the right displays the according information. B. HumanMachine Interaction The base of the human machine interaction is a live view of the tablet’s camera in which the sticky notes are overlaid in an Augmented Reality fashion. When touching a note, a form with additional information pops up, see Figure 3. From a technologically viewpoint this is a HTML overlay, reflecting the client-server based architecture of the software. Beside the standard information for a sticky note like name, type, and description an additional tagging system has been implemented for a flat hierarchy of the notes. To each note one ore more pre-defined tags can be assigned, which allow an easy to use filtering, e.g.: showing only electronic related notes in the view. Examples of the tags can also be seen in Figure 2 and 3 on the bottom right. Furthermore the system supports different user roles, which differ in the amount of information they can edit and/or which sticky notes they can see. IV. TECHNICAL EVALUATION The system is currently in an evaluation phase. Tests regarding positioning showed up accuracies of±0.5cm in a volume of 5m×5m×5m. The used hardware is a standard Microsoft surface tablet with the internal camera set to a resolution of 1280×720px andamaximumachievable framerateof25fpsof theoverlay. However the limiting factor for the frame rate is the tablet camera itself. With an external industrial camera frame rates up to 100fps have been achieved. For the registration step and the moveable parts also the marker size and camera-marker-distance have been evalu- ated: for a reasonable marker size of 7cm the largest distance to the marker is 2m. With higher distances the detection and therefore the visualization becomes in-stable. V. CONCLUSION AND FUTURE WORK In this paper we presented an industrial grade AR based documentation system which extends the current state-of- the-art by integrating additional tracking modalities and furthermore allows the user to edit the information in an intuitive way. The proposed system is currently on a tech- nology readiness level of 7. The next steps are extensive field tests. Although the system can be used stand-alone, the ideal synergy is together with a PLM system. Currently the import/export functions support only one such system, however this will be extended in future. Furthermore of vital interest is the evaluation of time savings gained by using our system. For this a detailed study will be set up together with industrial partners. From an application point of view a possible extension could be the replacement of the PLM input with a Virtual Reality (VR) system. This would allow one user to add information on a model in VR which is then immediately shown to a different user on the real object. This could be the base for numerous multi-user scenarios e.g. the usage as remote maintenance system. The advantage over a traditional 2D camera assisted remote maintenance system would be the exact 3D positioning of the information on the object of interest. ACKNOWLEDGMENT This research is funded by the project SIAM (FFG, 849971) and by the European Union in cooperation with the State of Upper Austria within the project Investition in Wachstum und Bescha¨ftigung (IWB). REFERENCES [1] RE’FLEKT. Kothes / Docufy - Virtually enhanced handbooks with AR intergrated into your editing system, 2016. https://www.re-flekt.com/portfolio-view/ augmented-reality-technical-documentation/. [2] S. Garrido-Jurado, R. Mun˜oz Salinas, F.J. Madrid-Cuevas, and M.J. Marı´n-Jime´nez. Automatic generation and detection of highly reliable fiducial markers under occlusion. Pattern Recogn., 47(6):2280–2292, June 2014. [3] S. Akkaladevi, M. Ankerl, C. Heindl, and A. Pichler. Tracking multiple rigid symmetric and non-symmetric objects in real-time using depth data. In 2016 IEEE International Conference on Robotics and Automation (ICRA), pages 5644–5649, May 2016. 108