Photogrammetric measurement process capability for metrology assisted robotic machining

• Photogrammetry measurement performance is evaluated for assisted robotic machining.
• Results show uncertainty can be >170 μm and systematic error is consistently higher.
• Robustness to short-term drift and operator technique is shown in simple tests.
• Challenges are error variance with part geometry and inconsistent data acquisition.

This paper documents the findings of experiments done to assess the capability of a photogrammetric measurement process for use in a metrology assisted robotic machining application. Capability is judged from the perspective of uncertainty and error across various part geometries over several days. The influence of operator technique on error and point acquisition ability for challenging geometry is also assessed. The process is found to be resistant to short-term error drift when operating in a controlled environment but systematic and random errors are demonstrated to be highly dependent on geometry. Operator influence on capability is found to be minimal when scanning freeform geometry, although this is unlikely for more complex parts. Acquisition of inspection points on cylindrical pockets is found to be a limitation. Technological development opportunities are highlighted in the context of the metrology assistance application considered. Overall, a thorough assessment of the measurement process capability is made and findings provide a quantification of current state, setting a base case for comparing research progress against.