Article View: sci.image.processing
Article #693Geometrical Distortion
From: Patrick Law
Date: Wed, 14 Oct 1992 18:56
Date: Wed, 14 Oct 1992 18:56
20 lines
1269 bytes
1269 bytes
One of my recent homework assignment is about geometrcal distortion correction. We model the distortion using second order equations. That is, the reverse address mapping is: u = a0 + a1 * x + a2 * y + a3 * x ^ 2 + a4 * x * y + a5 * y ^ 2; v = b0 + b1 * x + b2 * y + b3 * x ^ 2 + b4 * x * y + b5 * y ^ 2; The idea is actually came from Pratt's "Digital Image Processing", page 430. A couple of points are picked and the unknowns a's and b's are solved by pseudo inversing the matrix. However, the a's and b's I obtained are very sensitive to the way that the reference points are picks. My result, of course, is unsat- isfactory. I wonder if anybody out there has done similar operation and could give me some hints or references. Many thanks. Patrick Law
Technical Details
Message-ID:
<1bjbo3INN6p9@girtab.usc.edu>Path:
arc1-ca.pugleaf.net!sparky!uunet!ogicse!uwm.edu!rpi!usc!chaph.usc.edu!girtab.usc.edu!not-for-mailShow Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!ogicse!uwm.edu!rpi!usc!chaph.usc.edu!girtab.usc.edu!not-for-mail From: plaw@usc.edu (Patrick Law) Newsgroups: sci.image.processing Subject: Geometrical Distortion Keywords: Pin-cushion, barrel distortions. Message-ID: <1bjbo3INN6p9@girtab.usc.edu> Date: 14 Oct 92 18:56:35 GMT Article-I.D.: girtab.1bjbo3INN6p9 Sender: plaw@girtab.usc.edu Organization: University of Southern California, Los Angeles, CA Lines: 20 NNTP-Posting-Host: girtab.usc.edu