If you know the desired Yaw and Pitch values for the image you wish to move then enter the data manually.

Last edited by mediavets (2009-01-13 10:10:51)

I don't know how you are currently shooting your 800-plus image sets?

You may be interested in the low cost Merlin/Orion+Papywizard robotic pano head system:
http://www.autopano.net/forum/p32070-20 … -37#p32070

This system can fully automate the shooting of hires, large mosaic panos. It also (optionally) creates a record of the Yaw/Pitch values of every shot taken which is stored in an XML format file on the Papywizard software host system:
http://trac.gbiloba.org/papywizard/wiki … putXMLfile

APPV2/APG has an import filter specifically for Merlin/Orion+Papywizard images sets, this reads the shooting position data in the XML file to help position the images when stitching and should prevent any orphaned or mispositioned images.

The Papywizard software that controls the Merlin/Orion pano head will run without connection to the head so you can download and try it to get a feel for how it works and what it can do - try the latest development version for Windows:
http://trac.gbiloba.org/papywizard/#Dev … entversion

Last edited by mediavets (2009-01-13 10:47:54)

panoperson wrote:

I think we already had this conversation.  I am using the AutoMate motorized pano head from The Gadget Works.  I don't get the pitch and yaw data from it, but I love the fact that it only weighs 2 pounds and runs all day on 4 AA batteries.

OK - sorry.

Then I guess you will have to live with the resulting limitations of your chosen pano head system.

OTOH if you are shooting a regular matrix it should not be that difficult to manually enter Y/P values for orphaned images - it may take some time of course if you have a lot of them.

Or maybe Frederic can modify Papywizard to add support for your Automate head?

Or maybe the GigaPan robot for DSLRs - said be be in development - will suit you?

Or maybe the generic row/column mosaic import filter to be developed for APPV2/APG will meet your needs?

Or maybe you can get Automate's controller modified to emulate the shooting sequence used by the GigaPan robot and then use the Gigapan import filter?

Or write a script to rename your images so that that they match the sequence used by the GigaPan robot and then use the GigaPan import filter?

Last edited by mediavets (2009-01-13 11:23:36)

fma38 wrote:

Do you have a link on this device?

http://www.thegadgetworks.com/

http://www.outbackphoto.com/CONTENT_200 … index.html

http://www.thegadgetworks.com/pdf/UserManual.pdf

I think it's based a highly modified version of these devices?:
http://www.b-hague.co.uk/pan__tilt_power_head.htm

http://www.21best.com/21_best/electroni … sale_.html

See also this thread:
http://www.autopano.net/forum/p33165-20 … -21#p33165

The Automate control program running on a Windows Mobile PDA appears to generate a program/script that is transmitted to the Automate head (via Bluetooth) where it is stored and executed on the head. A rather different concept from Papywizard. OTOH it would appear that it should be possible for the control program also to generate a Papywizard-style XML data file since it must compute shooting positions.

Last edited by mediavets (2009-01-13 11:47:39)

Yes, true about the NPP. But I don't understand how they can provide such accuracy without position feedback... Are you sure there are no encoders? Or they are using stepper motors, not DC motors...

BTW, how do you store your 1200 images? Do you periodically pauses the shooting to change the card? Or are you using a PC?

Some tips for you:
1- Tick "Force every image to be in the same panorama" when stitching such large panos.

2- Now that all your images are in this pano, click on "Guess unlinked image location" in the CP Editor. It works wonderfully on images that are missing from from between other images (not on the sides)!
(Please report back how well it worked for you, and show a screenshot before/after while you have the cp editor open. But do this only if you detected your pano using "force every image to be in the same panorama")

3- If you need to manually move an image, do what mediavets said - click on it and use yaw/pitch values. If you have trouble selecting the right one, hover your mouse over it and use the scrollwheel, it will flip between all images under the cursor.

panoperson wrote:

What I would really like to be able to do is specify all the images it overlaps and then instruct the program to stitch it as best it can to those images

4- You can do this. Once you get your image more or less into place (step 2 or 3), open the CP Editor, go back to the Pano Editor while keeping the CP Editor window open, rightclick on the moved unstitched image, and select "Local geometric analysis", then when it finishes griding (might take a while), rightclick again and select "Local optimize". Optimize colors afterwards. More explained in the wiki article Moving source images, steps 3 and 4 are for you.

fma38 wrote:

Yes, true about the NPP. But I don't understand how they can provide such accuracy without position feedback... Are you sure there are no encoders? Or they are using stepper motors, not DC motors...

BTW, how do you store your 1200 images? Do you periodically pauses the shooting to change the card? Or are you using a PC?

In AutoMate 1.0, there was a calibration procedure to set the number of milliseconds per degree.   In AutoMate 2.0 there is a much simpler system for doing this that only takes a few seconds,  but the result is the same - computing time per degrees.  You enter your focal length, aspect ratio, crop factor, and camera orientation to get a trial value and then tell it to shoot two overlapping images that you inspect in case you need to tweak the settings to get the desired overlap.  Then you can save your settings by name so that the next time you use the same camera and lens, all you have to do is reload the setttings.  It also always remembers the last settings you used so that the next time you start up, it uses those.   So unless you change camera or lens, you don't even have to reload anything.

I am using a 16 GB card in the camera and am shooting JPG format right now, so there is plenty of room.  But the software allows you to pause at any time to change cards.

DrSlony wrote:

Some tips for you:
1- Tick "Force every image to be in the same panorama" when stitching such large panos.

2- Now that all your images are in this pano, click on "Guess unlinked image location" in the CP Editor. It works wonderfully on images that are missing from from between other images (not on the sides)!
(Please report back how well it worked for you, and show a screenshot before/after while you have the cp editor open. But do this only if you detected your pano using "force every image to be in the same panorama")

3- If you need to manually move an image, do what mediavets said - click on it and use yaw/pitch values. If you have trouble selecting the right one, hover your mouse over it and use the scrollwheel, it will flip between all images under the cursor.

panoperson wrote:

What I would really like to be able to do is specify all the images it overlaps and then instruct the program to stitch it as best it can to those images

4- You can do this. Once you get your image more or less into place (step 2 or 3), open the CP Editor, go back to the Pano Editor while keeping the CP Editor window open, rightclick on the moved unstitched image, and select "Local geometric analysis", then when it finishes griding (might take a while), rightclick again and select "Local optimize". Optimize colors afterwards. More explained in the wiki article Moving source images, steps 3 and 4 are for you.

Great tips!   Thanks a lot.  I will try these out later today.

DrSlony wrote:

Some tips for you:
1- Tick "Force every image to be in the same panorama" when stitching such large panos.

Unfortunately, this causes APP to hang with the Optimize progress bar 1/3 the way across.

You could uncheck "strong algorithm" if that pano is huge, should speed things up.

Arrrghhh!  I worked for hours on moving images around, doing local geometric analyses, optimizing, etc.  only to get a critical error when moving an image and then watch the application close down on me.  It is too frustrating for words.  Maybe it is time to give PtGUI a crack at this pano.

Of course!  I should have realized that saving the project periodically would retain all my changes as I went.  I never even saved it after that long detect phase.  (Doh!) So what is the difference between Save and Extended Save?