Abstract

Augmented reality is used increasingly often in industrial assembly and maintenance use cases. However, most companies do not take full advantage of the fact that the CAD models of products are already available and limit themselves to the less sophisticated tracking approach of marker-based tracking. This thesis investigates the possibility of incorporating model tracking into a real industrial use case. An application was being developed that aids its operator to perform a simplified quality assurance procedure. High accuracy is required for the application to run continuously using only model tracking, even after losing visual contact with the target. Combining the model tracking capabilities of Vuforia Engine with AR Foundation’s more precise camera pose estimation is sufficient for such accuracy. Apart from this, the developed software is lacking in some areas that are non-critical to the research question, and thus, subsidiary. As such, the UI and workflow are neither sufficiently developed nor tested to be used in a real situation in their current forms. The tracking solution used in this thesis is a viable approach for many use cases and its complexity should not be perceived as daunting. Companies that try to innovate their production lines using AR applications should investigate whether existing CAD models can improve this concept.

Results/Implementation/Project Description

Conclusion

[ PDF (optional) ] 

[ Slides Kickoff/Final (optional)]