Methodology to Evaluate the Performance of Portable Photogrammetry for Large-Volume Metrology

Abstract

The increased relevance of large-volume metrology (LVM) in industrial applications entails certain challenges: measurements must be cost-efficient and the technologies must be easy to use while ensuring accuracy and reliability. Portable photogrammetry shows great potential to overcome such challenges, but industrial users do not yet rely on its accuracy for large scenarios (3 to 64 m), especially when mass-market cameras are not conceived of as industrial metrology instruments. Furthermore, the measurement results might also depend on the operator’s skills and knowledge of the key process variables. In this work, a methodology was designed so that the measurement uncertainty of portable photogrammetry can be evaluated under controlled conditions for LVM. To do so, PTB’s reference wall, which was designed to assess laser-based methods applied to large volumes, was used as a reference artefact to study the measurement performance under different conditions, enabling an analysis of the relative influence of two process variables: the spatial arrangement of the optical instruments on the scene, and the relative camera poses for an accurate triangulation. According to these variables, different measuring conditions were designed (Monte Carlo analysis), and experimentally evaluated and reported (LME, length measuring errors), analysing the performance figures expected from both unskilled and expert users.

Keywords

• large-volume metrology (LVM)
• portable photogrammetry
• path planning
• inline measurement