NZcaver wrote:Very impressive, Footleg!
What camera body and lens are you using, and how do you do your post processing?
This was taken with a Canon EOS 5D, which has a 12Mp full frame 35mm sensor. I used a 20mm prime lens, which enables me to cover one 360 degree rotation in 8 shots. I find it easier to picture where all the images fit in my head when there are 8 in a rotation, as you can imagine them looking at the 8 points of a compass (N, NE, E, SE, S, SW, W, NW). I am using a Nodal Ninja 3 panoramic mount on the tripod to eliminate parallax problems with image alignment. I take 8 images looking level (straight ahead) for the 8 points of the compass. Then another 8 with the camera looking up 60 degrees, another 8 with the camera looking down 60 degrees, then a couple straight up and a couple straight down at the very end without the tripod so that this can be removed from the final panoramic image. I also took a specific image of the model to ensure I had him all in a single good frame with the bear. Lighting was done using home made LED lights (the ones I gave a talk on the design of at the ICS2009/NSS convention in Texas). See my post here for plans:
viewtopic.php?f=4&t=10090#p87654 Stitching can be done with the free open source Hugin application. But in this case I used Autopano Pro. I did have to manually edit some control points in Hugin and then hand edit the Autopano file to insert these however, as I cannot find a way to manually add control points directly into Autopano. Some of the individual images were hand blended composites of two frames taken in the same position in order to even out lighting and get compatible shadows on the overlapping region with adjacent frames. Currently the image is the auto blended result, with the blender in Autopano given some guidance by masking out areas of some frames so that they would not be used (to force the best exposed and in focus parts of overlapping frames to be used, and to ensure the caver did not have parts of his body replace with the cave background from other frames!). I plan to hand blend some areas from the individual frames to further improve the image. But that will take some more hours and so far this image has taken many hours at the computer already!