Share Email Print

Proceedings Paper

Detection And Characterization Of Flaws On Machined Metal Surfaces
Author(s): Reza Safabakhsh; Rafael C. Gonzalez
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

This work studies detection and characterization of flaws on machined metal surfaces based on the flaw depth. Diffracted light resulting from the incidence of a laser beam on the metal surface is sensed by a solid-state detector and analyzed by a digital computer to determine the surface conditions. Statistical analysis and pattern recognition algorithms are used to find data features which strongly correlate with the flaw depth. Trainable classifiers use these features to detect surface flaws whose (maximum) depth exceeds a given threshold. The range of flaw depth and a rough depth estimate are also determined and problems arising due to detector characteristics or interaction of flaw parameters are discussed. Experimental results of near 100% correct detection and 50% to 85% correct range determinations are reported.

Paper Details

Date Published: 16 January 1989
PDF: 8 pages
Proc. SPIE 0954, Optical Testing and Metrology II, (16 January 1989); doi: 10.1117/12.947629
Show Author Affiliations
Reza Safabakhsh, Tennessee State University (United States)
Rafael C. Gonzalez, University of Tennessee & Perceptics Corporation (United States)

Published in SPIE Proceedings Vol. 0954:
Optical Testing and Metrology II
Chander Prakash Grover, Editor(s)

© SPIE. Terms of Use
Back to Top