Talk:Image rectification

Parallograms in Part 1
It would be a more informative diagram if the square boxes in the non-rectified Part 1 of the diagram were parallelograms (in the horizontal dimension) that translated to the (correctly) rectified square boxes in Part 2. See http://en.wikipedia.org/wiki/Eigenvector and http://en.wikipedia.org/wiki/File:OrthoPerspective.svg for examples that also explain what I mean especially the first. Regards 123.208.145.187 (talk) —Preceding undated comment added 13:11, 2 March 2010 (UTC).

Rectified image
The term "rectified image" is used to describe a transformation applied to a photograph of the lunar limb to correct for foreshortening. However, it doesn't appear to fit with the definition on this page. Are the two entirely different or can the concepts be merged?&mdash;RJH (talk) 23:28, 13 January 2011 (UTC)

"search in two-dimensions"
The article says "In most camera configurations, finding correspondences requires a search in two-dimensions." Even in unrectified images the search is one-dimensional, the search line is just not horizontal but slanted. Two-dimensional search would be the optical flow problem (correspondence search in non-static scenes or with unknown camera parameters). I can correct this shortly, if you want me to. — Preceding unsigned comment added by Lemonyfresh (talk • contribs) 05:19, 15 December 2017 (UTC)

"two or more images"
Rectification in geo information systems is possible for >2 images (I think) but in computer vision it's not. R. Szeliski "Computer Vision -- Algorithms and Applications" Section 11.1.1: "Note that in general, it is not possible to rectify an arbitrary collection of images simultaneously unless their optical centers are collinear, ..." I suggest removing "two or more" from the article's introduction and adding a short note to the article's computer vision subsection. — Preceding unsigned comment added by Lemonyfresh (talk • contribs) 05:30, 15 December 2017 (UTC)

Example
The given example is not clear. Especially when explains the transformations to apply. — Preceding unsigned comment added by 138.250.126.40 (talk) 16:11, 22 January 2020 (UTC)

External links modified (January 2018)
Hello fellow Wikipedians,

I have just modified 2 external links on Image rectification. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20100821194542/http://www.umiacs.umd.edu/users/sernam/papers/rect.pdf to http://www.umiacs.umd.edu/users/sernam/papers/rect.pdf
 * Added archive https://web.archive.org/web/20110706161552/http://www.matmidia.mat.puc-rio.br/sibgrapi2009/media/undergraduate_work/60067.pdf to http://www.matmidia.mat.puc-rio.br/sibgrapi2009/media/undergraduate_work/60067.pdf

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 06:57, 26 January 2018 (UTC)

Improved rectification algorithm
I've suggested to link this new rectification algorithm, which is an improvement of the one by C. Loop, already linked in the article. The algorithm offers a closed form solution for rectification and code is also available. However, my changes have been reverted and I am not sure which rule I have misinterpreted. Cheers. 138.250.148.250 (talk) 14:45, 26 September 2022 (UTC)
 * Well, we are an encyclopedia, and the purpose of our articles is really to have text, verified by secondary sources. You dropped a few links in there, and we're not a link dump. Drmies (talk) 01:59, 27 September 2022 (UTC)
 * Before your edit there was a "reference implementations section" that was doing exactly that. No one dumped links.
 * Moreover, I don't get why some non-optimal and very old solutions like Fusiello get to stay in the references, while new ones (described by a published paper) cannot be added. 138.250.239.12 (talk) 09:09, 27 September 2022 (UTC)