This preview shows page 1. Sign up to view the full content.
Unformatted text preview: on phase to address variability limits are in biunivocal relationship with the sets of operations used in the verification phase to identify the element subject to the specification and to evaluate its conformity. 2. THE IDENTIFICATION OF PRODUCT SHAPE The paper proposes a probabilistic approach to the recognition, segmentation and characterization of geometrical shapes based on a data-driven methodology applied to a set of measured points. The use of statistical empirical tools offers many advantages in term of clearness, reliability and simplicity thus providing a unifying approach to the geometrical control of product shape. As a matter of fact, statistical control techniques are widely adopted in verification and manufacturing phases and their extension to the design phase could improve quality and reduce costs. Other interesting methodologies [Fischer et al., 2003], [Gelfand et al., 2004] are mainly focused on the inspection phase and do not achieve a mathematical universal description of 3D shapes which could be adopted in the verification phase and in the design phase too. Moreover, these methodologies do not exploit the advantages of the duality 172 P. Chiabert and M. De Maddis principle (thus preventing a direct matching of inspection results...
View Full Document
- Spring '10
- The Land