Modern coordinate measurement machines have provided industry with new tools for inspecting complex parts. This paper develops statistical procedures that complement these inspection methods, and it ties the conformance problem into the hypothesis testing formalism of conventional statistics. Also, the paper suggests strategies for efficient sampling of the part surface, and an implementation of the usual decision-theoretic formulation of the tradeoff between false acceptance and false rejection of part geometries. These ideas are illustrated through the analysis of real and simulated data for perfect and imperfect cylinders.

This content is only available via PDF.
You do not currently have access to this content.