DocumentCode
2576370
Title
Constraints on quadratic curves under perspective projection
Author
Safaee-Rad, R. ; Smith, K.C. ; Benhabib, B. ; Tchoukanov, I.
Author_Institution
Toronto Univ., Ont., Canada
fYear
1991
fDate
13-16 Oct 1991
Firstpage
57
Abstract
The authors address the problem of three-dimensional (3-D) location estimation based on quadratic-curved features. They derive the mathematical relations or constraints on the 3-D position and orientation of quadratic-curved features using the standard rotation and the standard transformation concepts introduced by K.I. Kanatani, (1988), and assuming that the true size and shape of a given quadratic feature are known a priori, with its projection image given. In this context, an analytical method is introduced for estimation of the standard rotation and determination of the shape of a quadratic-curved feature at its canonical position. It is shown that, in general, knowledge of the true shape and size of a quadratic-curved feature does not yield a sufficient number of constraints to determine the 3-D position and orientation uniquely. As a result, extra constraints must be acquired from various sources of information and fused with these constraints to obtain unique 3-D position and orientation estimates
Keywords
optimisation; parameter estimation; pattern recognition; 3D location estimation; 3D position estimation; constraints; orientation estimates; pattern recognition; perspective projection; projection image; quadratic-curved feature; rotation concept; transformation concepts; Computer integrated manufacturing; Computer science; Computer vision; Educational institutions; Laboratories; Layout; Machine vision; Mathematics; Mechanical engineering; Shape;
fLanguage
English
Publisher
ieee
Conference_Titel
Systems, Man, and Cybernetics, 1991. 'Decision Aiding for Complex Systems, Conference Proceedings., 1991 IEEE International Conference on
Conference_Location
Charlottesville, VA
Print_ISBN
0-7803-0233-8
Type
conf
DOI
10.1109/ICSMC.1991.169661
Filename
169661
Link To Document