TY  - SER
AU  - Bukhari,Faisal
AU  - Dailey,Mathew N.
AU  - Afzulpurkar,Nitin V.
AU  - Duboz,Raphael
ED  - Higher Education Commission (HEC), Pakistan,
ED  - Asian Institute of Technology Fellowship,
TI  - Automatic radial distortion estimation from a single image
T2  - Dissertation
PY  - 2012///
CY  - Pathum Thani, Thailand
PB  - Asian Institute of Technology
KW  - Computer vision
KW  - Image processing
N1  - A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Computer Science, School of Engineering and Technology; Thesis (Ph.D.) - Asian Institute of Technology, 2012
N2  - Many computer vision algorithms rely on the assumptions of the pinhole camera model, butlens distortion with off-the-shelf cameras is usually significant enough to violate this as-sumption.  Many methods for radial distortion estimation have been proposed, but they allhave limitations.  Robust automatic radial distortion estimation from a single natural imagewould be extremely useful for many applications, particularly those in human-made envi-ronments containing abundant lines.  For example, it could be used in place of an extensivecalibration procedure to get a mobile robot or quadrotor experiment up and running quicklyin an indoor environment.In this dissertation we propose a new and fully automatic method for radial distortion esti-mation based on the plumb-line approach.First, the method works from a single image and does not require a special calibration pattern.It is based on Fitzgibbon{u2019}s division model.Second, we devise a new algorithm for robust estimation of circular arcs.Third,  we design and implement a new algorithm for robust estimation of lens distortionparameters based on the estimated circular arcs.Fourth, we perform an extensive empirical study of the method on synthetic images.  Wedevelop our own data set for synthetic images under different levels of lambda and distortioncenter.Fifth, we perform a comparative statistical analysis of how different circle fitting methodscontribute to accurate distortion parameter estimation.Sixth,  we  provide  qualitative  results  on  a  wide  variety  of  challenging  real  images.   Theexperiments demonstrate the method{u2019}s ability to accurately identify distortion parametersand remove distortion from images. Seventh, we perform a direct comparison of our methodwith that of Alvarez et al. (Alvarez, Gomez, & Sendra, 2009), the only researchers who haveprovided a publicly accessible implementation of their method, on synthetic images.Finally,  we  provide  the  source  code  based  on  OpenCV  (Bradski,  2000)  online1for  re-searchers interested in evaluating or extending our procedure
UR  - http://203.159.5.9/ait-thesis/detail.php?q=B00292
ER  -