M. S. Thesis Abstract

A Methodology for Generating Physically Accurate Synthetic Images for Machine Vision Applications

Johné Michelle Parker

March 1992

The design of a computer vision system for part presentation is a complex problem.  To design an efficient low-cost system, the interdependence of the system hardware and software must be ocnsidered.  Numerical simulation can be a powerful aid in examining the performance of an integrated hardware/software system.  However, numerical simulation requires a method to generate physically accurate synthetic images that simulate the images captured by the vision system.  This paper describes a methodology for generating the required physically accurate images.  A comparison of synthetic images to images captured using an existing machine vision system is also done.