I've been having a go at rectifying some images with the Geothings Warper and I've covered about half of Tiddington. They're public on the website as numbers 1310, 1307, 1309, 1312, 1313, 1311, 1308 (in that order, most vertical first). However, there are very obvious seams between images. This got me thinking about the possibility of creating a panorama image to actually stitch these images together.
Hugin is an open source panorama program (which I believe uses parts of the SIFT algorithm to match features) which can do a very good job of stitching the images we've collected together. I had a practice with images 00822-00825 and very successfully was able to stitch them together with no visible seam at all. I then uploaded the composite image to Warper (http://warper.geothings.net/maps/1315 or the original higher-quality image at http://warper.geothings.net/uploads/1315/original/test3.jpg) and rectified it. However, the recfication in this case probably isn't good enough due to the slight warping caused by Hugin. What we really need is a Hugin-type application which allows you to do the standard stitching thing (setting control points between photos) while at the same time allowing you to match points on the photos against points on a map such that when it creates the 'panorama', it's automatically rectified too. Does anyone know of a way to utilise Hugin or Panorama Tools (upon which Hugin is based) to do this sort of thing? I imagine it's not been done exactly like this before but I'm sure there must be a way. The long way to do it (and I'm not really suggesting it) would be to warp each individual image in Warper, then stitch the warped images with Hugin (at this point they should almost match the map) and then finally re-Warp them again to neaten them up. The problem with this is that Warper only produces low resolution images which wouldn't work with Hugin very well. Thoughts/suggestions? -- Matt Williams http://milliams.com _______________________________________________ Talk-GB mailing list [email protected] http://lists.openstreetmap.org/listinfo/talk-gb
