The practical implementation of nontrivial machine vision algorithms requires a balance of efficiency, flexibility, and cost. This paper discusses the development and implementation of an industrial system that recognizes scanned rectangular kitchen cabinet frames. By utilizing a configuration of modular algorithms that reduce the image into structural primitives, accurate recognition is possible at a relatively high speed using limited hardware. The system essentially decomposes a silhouette of the frame into a border representation, extracting corner-containing regions which, in turn, yield vertices. Global information about the frame is then used to convert the vertices into usable features. This thesis discusses the motivation, development, and implementation of this system. Recognition tests were performed successfully on several thousand frame samples. / Master of Science
Identifer | oai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/40636 |
Date | 17 January 2009 |
Creators | Del Gigante, Christopher John |
Contributors | Electrical Engineering, Abbott, A. Lynn, Athanas, Peter M., Bay, John S. |
Publisher | Virginia Tech |
Source Sets | Virginia Tech Theses and Dissertation |
Language | English |
Detected Language | English |
Type | Thesis, Text |
Format | v, 122 leaves, BTD, application/pdf, application/pdf |
Rights | In Copyright, http://rightsstatements.org/vocab/InC/1.0/ |
Relation | OCLC# 36118298, LD5655.V855_1995.D454.pdf |
Page generated in 0.0015 seconds