3-D body mapping with computer vision

The potential and accuracy aspects of a completely automated stereo photogrammetric vision system for performing precise 3-D body mapping of whole human bodies have been investigated. A prototype vision system by integrating structured lighting, computer vision, and stereophotogrammetric techniques was developed and implemented to map an entire human body in a highly automatic manner. The system consisted of nine CCD cameras arranged in three stereo triples, a 486-based personal computer equipped with two frame grabbers, and three slide projectors for projecting dot patterns on the body. A movable control field was developed for CCD camera calibration on site. A method was developed for the automatic and simultaneous calibration of the interior and exterior orientations of all cameras. Algorithms were developed for the automatic measurement of image coordinates, for finding corresponding matching points in a stereo triple, for integrating multiple stereo models, and for computation of volumes, surface areas, and circumferences. The stability of the prototype system during the system warm-up period was studied and the potential accuracy and limitations of the 3-D body mapping with computer vision were evaluated by analyzing the results of mapping four human subjects.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Civil Engineering

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2011

Author dc:creator

• Ke, Ying

Contributors dc:contributor

• Wong, Kam W.
• Frank, Thomas D.
• Huang, Thomas S.
• Slaughter, Mary H.

• Engineering, Biomedical
• Engineering, Electronics and Electrical
• Remote Sensing

Statement dc:rights

• Copyright 1995 Ke, Ying

Language dc:language

eng

Identifier

AAI9543622
(UMI)AAI9543622

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/23328