Premium
On vision systems identification with application to fixed‐camera robotic systems
Author(s) -
Kelly Rafael,
Reyes Fernando
Publication year - 2000
Publication title -
international journal of imaging systems and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.359
H-Index - 47
eISSN - 1098-1098
pISSN - 0899-9457
DOI - 10.1002/1098-1098(2000)11:3<170::aid-ima1002>3.0.co;2-l
Subject(s) - computer science , computer vision , artificial intelligence , frame (networking) , camera resectioning , identification (biology) , distortion (music) , rotation (mathematics) , property (philosophy) , machine vision , least squares function approximation , set (abstract data type) , calibration , mathematics , telecommunications , amplifier , computer network , philosophy , botany , statistics , bandwidth (computing) , epistemology , estimator , biology , programming language
Vision system calibration and identification are important issues for effective implementation of high‐performance robotic systems. Vision system identification addresses the problem of determining the mapping from points in the world frame to their corresponding location in a computer image frame. By assuming rotation of the camera frame around one of the principal axes of the world frame—but incorporating radial lens distortion—we show that this mapping can be expressed as a linear regression model in terms of a suitable combination of the intrinsic and extrinsic camera parameters. This property allows the application of several known techniques based on resolution of a determined set of linear equations and least‐squares–based methods to estimate these parameters from experimental input‐output data. Experimental comparisons are carried out to illustrate the performances of these methods. © 2001 John Wiley & Sons, Inc. Int J Imaging Syst Technol 11, 170–180, 2000