 |
|
|
|
|
|
 |
| ||||||||
 |
 
|
|
|
|
|
|
|
|
|
| ||||||||
|
|
| 
| User list | Rules | You are not logged in.
Using a motorized panohead it would be nice to have such an option: I guess it'd bring to more precise results.
Also, for spherical pano, an independent lock: that is, lock roll for all images and lock pitch only for the around images (or by row, in case of hi-res spherical panos)
How can I do this?
If still not possible, is it something we could see on future versions?
Thanks,
Offline
marco.lanciani wrote:
How can I do this?
You can't.
As I undersatnd it - and I may be quite mistaken - unless you use one of the robotic pano head import filters, or a template created from previous stitch, then APP/APG 'knows' nothing about the absolute or relative orientation, nor about the relative positioning of images, it just searches for matching features.
Offline
If so, I'd be interested in such an option because I use a robotic panohead and an import preset... looking forward to see it!
Offline
marco.lanciani wrote:
If so, I'd be interested in such an option because I use a robotic panohead and an import preset... looking forward to see it!
If so what do you expect such an option would do for you?
Which robotic head do you use?
Offline
- I expect a better stitching right on first detection or at least better (closer to the truth) lens distortion values and final panorama dimensions.
- Merlin
Offline
marco.lanciani wrote:
- I expect a better stitching right on first detection or at least better (closer to the truth) lens distortion values and final panorama dimensions.
- Merlin
If you are using Merlin/Panogear and the Papywizard Import filter then APP/APG already has relatively precise pitch and yaw data yet Kolor choses not to lock these values.
Presumably that doesn't produce better results with their detection algorithmns.
As I understand it they make use of 'information' made available via the various robotic pano head import filters - be that relatively precise position data, or just the simple dimensions of a regular grid in the case of the Gigpan heads - to constrain the search for matching features solely to neighbouring images, rather than searching all images for matching features with all other images.
Last edited by mediavets (2012-12-06 20:28:12)
Offline
mediavets wrote:
If you are using Merlin/Panogear and the Papywizard Import filter then APP/APG already has relatively precise pitch and yaw data yet Kolor choses not to lock these values.
My doubts are not about Yaw data; only Pitch and Roll. Speaking in theory:
- Pitch is exactly the same on a row and optimizing for the Pitch on the same row might lead to wrong values on lens distortion parameters and so a slightly altered panorama dimension.
- Same thing, if not worse, for the Roll which is the same for the whole panorama because the camera doesn't rotate on its axis during exposures.
mediavets wrote:
Presumably that doesn't produce better results with their detection algorithmns.
Maybe the APG team already did some test in this direction: if so, it would be nice to hear from them.
By the way, it'd be nice to have the option to lock pitch and roll either. 
Offline
marco.lanciani wrote:
- I expect a better stitching right on first detection or at least better (closer to the truth) lens distortion values and final panorama dimensions.
- Merlin
The xml positions the images in the first step. Because of no head in the world is that prescise to position a camera pixel-perfect there needs to be optimization after the import.
If you have trouble with that check "skip optimization" in the xml import. Use manual optimization after the import resp. set hard-links between positions without CPs.
That can be a tricky thing to do . . . 
I guess there should be an option to set the radius inside which the optimizer looks for CPs - maybe it works be setting the sensitivity to low for being not too sensitive while optimizing. I didn´t find any description about it.
best, Klaus
Offline