Quality control tools using augmented reality
Objective:
Development of a quality control solution to be integrated into an AR-supported fabrication process. This solution will be based on computer vision and machine learning using smart glasses cameras.
Challenges:
- Limited embedded processing power of smart glasses.
- Integration within AR applications.
- Robustness of quality control under varying external conditions such as lighting.
Expected Working Principle:
Previous AACOMA Project Approach:
AR provides virtual instructions to the operator. Operator completes the task relying on expertise, then moves to the next step.
New SMARAGD Project Approach (BB1):
AR application assists the operator, and after task completion, the system checks for errors like foreign objects or incorrect placements. Operator can only proceed once quality test is passed.
Initial Solutions to be Investigated:
Direct use of smart glasses cameras to process images or videos.
- Evaluate the potential of Research Mode or OpenCV on Hololens for processing camera streams.
Vuforia Engine for image and model tracking.
- Detect component positions using 3D model and image tracking.
- Limited by its closed-source nature.
Remote computations to overcome processing power limitations.
- Share camera streams with a remote computer for processing and return the results to the smart glasses.
- Alternatively, use an external camera for remote processing, but this might face reference coordinate system issues.
