A fast algorithm for high accuracy gauging using computer vision - Robotics Institute Carnegie Mellon University

A fast algorithm for high accuracy gauging using computer vision

Morris Goldstein, Paul K. Wright, and David A. Bourne
Journal Article, Robotics and Computer-Integrated Manufacturing, Vol. 2, No. 2, pp. 105 - 113, December, 1985

Abstract

A vision-based inspection system capable of measuring convex-profile cylindrical objects has been developed. Both vertical and horizontal measurements are possible as well as cross-sectional reconstructions. A new algorithm has been developed that minimizes the number of camera views needed to obtain accurate reconstructions. This increases the speed of the analysis procedure. A production version of the inspection system has been installed in a forging plant; for this application camera lenses and other equipment dimensions give a measurement accuracy of ±1.0 mm on length and ±1.5 m2 on cross-sectional area.

BibTeX

@article{Goldstein-1985-121548,
author = {Morris Goldstein and Paul K. Wright and David A. Bourne},
title = {A fast algorithm for high accuracy gauging using computer vision},
journal = {Robotics and Computer-Integrated Manufacturing},
year = {1985},
month = {December},
volume = {2},
number = {2},
pages = {105 - 113},
}