Thanks Andreas On Dec 26, 1:16 pm, Andreas Metzler <[email protected]> wrote: > Tom Sharpless <[email protected]> wrote: > > On Dec 25, 1:41 pm, Yuval Levy <[email protected]> wrote: > >> On December 25, 2010 12:38:39 pm Tom Sharpless wrote: > >>> I keep suggesting that someone should make Hugin do CP finding > >>> intelligently, when the arrangement of the source images is already > >>> known. But so far nobody has volunteered. > >> when did you last check Hugin? Thomas Modes has added the > >> multi-row strategy a while ago... > > I have 2010.2.0.d8ce0ba947cc open in front of me. Like most other > > recent releases it is unable to generate any control points, and > > offers no help to understand why. It does present under preferences/ > > control point detectors a promising-looking option called "Autopano- > > SIFT-C (multrow/stacked)". But of course that is not supported by the > > files in the release. > > [...] > > Hello Tom, > > Upgrading does not automatically make the new control point detector > presets available. You'd need to either add them manually or use the > [reset to default] button. > > If you have autopano-sift-c available, choosing Autopano-SIFT-C > (multirow/stacked) as cp-detector instead of regular Autopano-SIFT-C > should give better (i.e. faster) results. > > The heuristics used is described > onhttp://wiki.panotools.org/Hugin_Parameters_for_Control_Point_Detector... > > Afaict there is no way for the user to simply change the way it works > (except manually for selecting a subset of images in which cps are to > be searched). The heuristics assume that images are sorted this way: > > 1 2 3 4 > 5 6 7 8 > ... > > If autopano-sift-c is not available to you the setting obvously will > not work for you. Especially in that case you really should upgrade to > 2010.4.0 rc1, which *includes* cpfind. >
Well, I configured the "Autopano-SIFT-C (multirow/stacked)" option to use the 2-step finder (from a previous build) and set it to work on a 20 x 2 array, shot with a 50mm lens. After 7 minutes 30 sec it reported 11 unconnected groups. One of those contained the complete upper row, several others were correctly aligned segments of the lower row (which is mostly water). Then I tried the "all images at once" option. It took about 13 minutes but produced a better result: 14 images in one connected group, plus six disconnected images. Under the same option, the one-step apsc crashed (out of memory) in the matching stage. PTGui aligned this set in 2 min 30 sec, leaving 4 disconnected images in the lower row. I will repeat once more and then shut up: Hugin's CP finding is a disgraceful mess, which prevents me -- and I expect many others -- from using this otherwise excellent stitcher. At the very least, there must be a web page that spells out how to get and use the various finders now available. I am sick of hearing "oh, that is patented, so we can't talk about it." Yes You Can. At best, someone will put together a distributable package that 'just works'. Best for 2011, Tom -- You received this message because you are subscribed to the Google Groups "Hugin and other free panoramic software" group. A list of frequently asked questions is available at: http://wiki.panotools.org/Hugin_FAQ To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/hugin-ptx
