Files in this item



application/pdf9114325.pdf (8MB)Restricted to U of Illinois
(no description provided)PDF


Title:Rigid object motion estimation from intensity images using straight-line correspondences
Author(s):Liu, Yuncai
Doctoral Committee Chair(s):Huang, Thomas S.
Department / Program:Electrical and Computer Engineering
Discipline:Electrical Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Engineering, Electronics and Electrical
Physics, Electricity and Magnetism
Computer Science
Abstract:Determining three-dimensional (3D) rigid object motion from intensity images is a central problem in computer vision. This thesis presents a computational approach of 3D rigid object motion estimation using straight line correspondences. Based on the rigidity assumption, a linear algorithm and a nonlinear algorithm of motion estimation are developed. With six or more line correspondences over three frames, the nonlinear algorithm first solves for the rotations, then eliminates the rotations from the original motion and computes the translations. The linear algorithm requires at least thirteen straight line correspondences over three frames to recover motion. A set of intermediate parameters is first estimated, and motion parameters are then determined from these intermediate parameters.
Also discussed in this thesis are techniques of straight edge extraction and straight line matching used for obtaining line correspondences of significant edges with high accuracy for motion estimation.
The algorithms of motion estimation have also been extended to two applications. The first extension considers the problem of camera location determination from lines and points, and the second, the problem of 3D rigid object motion estimation from image corners.
Experimental results on synthetic as well as real image data are given.
Issue Date:1990
Rights Information:Copyright 1990 Liu, Yuncai
Date Available in IDEALS:2011-05-07
Identifier in Online Catalog:AAI9114325
OCLC Identifier:(UMI)AAI9114325

This item appears in the following Collection(s)

Item Statistics