Approving Vision-Based Measurement Systems in the Presence of Within-Part Variation

ABSTRACT

Measurement systems analysis (MSA) techniques combine designed experiments and quantitative criteria to approve gauges. MSA techniques are predicated on a set of assumptions that may not all apply to some vision-based measurement system applications. This article presents a methodology for MSA under two conditions: the measuring device is robust to minor differences in how a part is oriented in the measuring fixture and the quality characteristic has within-part variability. The methodology includes the design for a gauge study, the analysis of variance (ANOVA) model used to fit the data, and estimators for the variances used as inputs to standard MSA criteria. The methodology is developed in the context of a vision gauge used to measure features that vary across the surface of a manufactured component yet generalizes to any measurement system with the two characteristics specified above. The techniques are demonstrated with manufacturer-provided data.

KEYWORDS:

• designed experiments
• gauge R & R
• measurement systems analysis
• vision systems
• within-part variation

Additional information

Notes on contributors

Karl D. Majeske

Karl D. Majeske is an Associate Professor of Quantitative Methods Management in the School of Business Administration at Oakland University in Rochester, Michigan. He received B.S., M.S., and Ph.D. degrees in industrial and operations engineering from the University of Michigan in 1985, 1991 and 1995, respectively. Dr. Majeske's research interests include measurement systems analysis, automotive body manufacturing, and financial and economic decision making.