DocumentCode :
2749062
Title :
An Analytic Solution of a Linear Camera Self-Calibration
Author :
Guo, Yang ; Fang, Zheng ; Xu, Xinhe
Author_Institution :
Dept. of Math., Northeastern Univ., ShenYang
Volume :
2
fYear :
0
fDate :
0-0 0
Firstpage :
9930
Lastpage :
9934
Abstract :
We propose an analytic solution for the self-calibration of five-parameter linear camera from three pairs of views of a scene, of four-parameter linear camera from two pairs of views and of two-parameter linear camera from one pairs of views based on Kruppa equations, respectively. The approach does not need any initial value, and can calculate accurate solution. The analysis was performed first by writing a suitable set of equations under introducing unknown scale factors from Kruppa\´s equations. Then, by an elimination scheme and symbolic computation, only one unknown scale factor was retained, so the equation set was reduced to a 30th order univariate polynomial equation for five-parameter or four-parameter camera, and a 3rd order univariate polynomial equation for two-parameter camera. After scale factors were attained, the camera intrinsic parameters were computed by a linear method. A numerical example was reported that confirms the theoretical results by "Mathematica" software
Keywords :
calibration; cameras; computer vision; matrix algebra; polynomials; symbol manipulation; Kruppa equations; Mathematica; Sylvester dialytic method; linear camera self-calibration; symbolic computation; univariate polynomial equation; Artificial intelligence; Calibration; Intelligent robots; Layout; Mathematics; Nonlinear equations; Optimization methods; Polynomials; Robot vision systems; Smart cameras; Analytic solution; Camera self-calibration; Kruppa’s equations; Sylvester’s dialytic method;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Control and Automation, 2006. WCICA 2006. The Sixth World Congress on
Conference_Location :
Dalian
Print_ISBN :
1-4244-0332-4
Type :
conf
DOI :
10.1109/WCICA.2006.1713938
Filename :
1713938
Link To Document :
بازگشت