Image-stitching and virtual tour solutions My account Updates
It is currently Sat Oct 25, 2014 8:46 am

All times are UTC + 1 hour




Post new topic Reply to topic  [ 9 posts ] 
Author Message
PostPosted: Sun Mar 09, 2014 7:46 pm 
Offline
Member
User avatar

Joined: Sun Oct 21, 2012 2:26 am
Posts: 41
Location: Paradise, California
I’ve attached a JavaScript web page program for generating a Papywizard XML output file for the GigaPan Epic 100.

One of the great features of Papywizard is its ability to generate an XML output file with all the image position information that can be fed into the stitching software to help orient the images before stitching. It’s especially helpful for panoramas with large areas of blue sky or indistinct clouds. Unfortunately, the Epic 100 does not provide for computer control so I’ve had to struggle with manually placing control points on images of featureless skies. Using Papywizard in simulation mode wasn't possible due to the way the Epic 100 works. The attached JavaScript web page that I created allows one to create a Papywizard XML output file for a wide range of panorama settings on the Epic 100.

It does require a bit of discipline, however, to record all of the settings you used at the time you take the panorama so that they can be fed into the program. Recording the vertical FOV used in the Epic 100 for each panorama is critical, it’s the only way the program can calculate the position of each image. I originally thought that I could devise a formula for calculating the horizontal and vertical offset based on the FOV. Unfortunately, although the offset is based on the FOV, the relationship is neither linear nor continuous. So I carefully measured the offsets for 16 different FOV values and built them into a look-up table. As a result, you are restricted to using one of the 16 values ranging from 3 degrees to 60 degrees. Set the Epic 100 FOV as you normally would except before clicking “OK” to the measurement, reduce it to the next lower value available in the calculator (e.g., if you measure 27 degrees, move the head so it reads 25 degrees and write it down).

Also, to get the most accurate position information you need to follow this procedure when setting a panorama with the Epic 100. Just before setting the right hand corner of the mosaic or 360 panorama make sure that the camera is at the midpoint of the last image by doing the following:
1. Record the indicated number of columns and rows. Pan the camera to the left so that the number of columns just changes to one less than the original indicated number and identify a landmark on the right edge of the image.
2. Pan the camera to the right so that the number of columns just changes to one more than the original indicated number and identify a second landmark on the right edge of the image.
3. Pan the camera back to the left until the right edge is at the midpoint between the two landmarks you identified.
4. Pan the camera up so that the number of rows just changes to one less than the original indicated number and identify a landmark on the bottom edge of the image.
5. Pan the camera down so that the number of rows just changes to one more than the original indicated number and identify a second landmark on the bottom edge of the image.
6. Pan the camera back up until the bottom edge is at the midpoint between the two landmarks you identified.
7. Click "OK" to finish defining the panorama range.

To use this program, save the attached .txt file to your computer and change the extension to ".htm". Open it with a web browser. You may have to click through a message regarding JavaScript in your browser if you have it blocked.
If you find a problem with this let me know and I’ll see if I can fix it.


Attachments:
Papywizard Mosaic Generator Ver 1.txt [20.79 KiB]
Downloaded 37 times
Top
 Profile  
 
PostPosted: Sun Mar 09, 2014 8:02 pm 
Offline
Member

Joined: Wed Nov 14, 2007 2:12 pm
Posts: 14045
Location: Isleham, Cambridgeshire, UK.
phill.butte wrote:
I’ve attached a JavaScript web page program for generating a Papywizard XML output file for the GigaPan Epic 100.

One of the great features of Papywizard is its ability to generate an XML output file with all the image position information that can be fed into the stitching software to help orient the images before stitching. It’s especially helpful for panoramas with large areas of blue sky or indistinct clouds. Unfortunately, the Epic 100 does not provide for computer control so I’ve had to struggle with manually placing control points on images of featureless skies. Using Papywizard in simulation mode wasn't possible due to the way the Epic 100 works. .


Why can't you use Papywizard in simulation mode?


Top
 Profile  
 
PostPosted: Sun Mar 09, 2014 8:27 pm 
Offline
Member
User avatar

Joined: Sun Oct 21, 2012 2:26 am
Posts: 41
Location: Paradise, California
The only information available from the Epic 100 when you set up a panorama is the vertical field of view of the image and the number of rows and columns. A key variable is the amount of horizontal and vertical overlap and that's not available from the Epic 100. I measured a vertical overlap of around 32% from 3 to 5 degrees FOV, 31% to 41% from 7.5 to 35 degrees and then 35% to 37% from 40 to 60 degrees FOV. I found it much more straightforward and reliable to just measure the angles each of the two axes move for a specific FOV value and use that in the calculation. Using Papywizard would mean basing the movements on the % overlap, which I think is unreliable in this situation.


Top
 Profile  
 
PostPosted: Sun Mar 09, 2014 8:33 pm 
Offline
Member

Joined: Wed Nov 14, 2007 2:12 pm
Posts: 14045
Location: Isleham, Cambridgeshire, UK.
phill.butte wrote:
The only information available from the Epic 100 when you set up a panorama is the vertical field of view of the image and the number of rows and columns. A key variable is the amount of horizontal and vertical overlap and that's not available from the Epic 100. I measured a vertical overlap of around 32% from 3 to 5 degrees FOV, 31% to 41% from 7.5 to 35 degrees and then 35% to 37% from 40 to 60 degrees FOV. I found it much more straightforward and reliable to just measure the angles each of the two axes move for a specific FOV value and use that in the calculation.

So do I understand that the Gigapan head uses different amounts of overlap between image s depending on the overal pano/scene VFOV?

If so, I wonder why Gigapan designed it like that?

Does it compute a similar amount of overlap on both axes?

Quote:
Using Papywizard would mean basing the movements on the % overlap, which I think is unreliable in this situation.

Is it unreliable in practice?


Top
 Profile  
 
PostPosted: Sun Mar 09, 2014 8:42 pm 
Offline
Member
User avatar

Joined: Sun Oct 21, 2012 2:26 am
Posts: 41
Location: Paradise, California
That's correct, a different amount for each FOV. I sent GigaPan support a request for information on the formula or lookup table they used but never got a response.

I can't say if its unreliable in practice, but % overlap is a derived value from my measurements of the angles. It might turn put to be "close enough" in practice but I think using the actual angles will give me more accurate results.


Top
 Profile  
 
PostPosted: Sun Mar 09, 2014 9:04 pm 
Offline
Member

Joined: Wed Nov 14, 2007 2:12 pm
Posts: 14045
Location: Isleham, Cambridgeshire, UK.
phill.butte wrote:
That's correct, a different amount for each FOV. I sent GigaPan support a request for information on the formula or lookup table they used but never got a response.

I can't say if its unreliable in practice, but % overlap is a derived value from my measurements of the angles. It might turn put to be "close enough" in practice but I think using the actual angles will give me more accurate results.


I would be interested in seeing a comparison.


Top
 Profile  
 
PostPosted: Sun Mar 09, 2014 10:49 pm 
Offline
Member
User avatar

Joined: Sun Oct 21, 2012 2:26 am
Posts: 41
Location: Paradise, California
I just ran a comparison on the two methods and here's what I found for the test pano I did yesterday. The pano was 10 columns by 5 rows with a Canon 60D and 200 mm lens. The image FOV was 5 degrees. The first problem is that the horizontal and vertical overlap figures for the Epic 100 are different for almost all FOV settings, sometimes by as much as 10%. In this case the vertical overlap is 33.2% and the horizontal 29.9%. Papywizard only allows one setting for overlap and it doesn't allow fractions, so I used 33%.
Papywizard calculated that the pitch head movement was 2.32 degrees and the yaw movement was 3.88 degrees for each image. The actual is 2.67 and 4.17. Papywizard calculated the total pitch movement at 11.6 degrees and the total yaw movement at 38 .8 degrees, vs. 13.35 and 41.67 actual. That's a total difference of 13.1% for the pitch and 7% in the yaw axis.
So I would say that for a small panorama using the Papywizard simulator for the Epic 100 would probably be "close enough". For the large, 300-400 gigapixel panos that I like to do I think the accumulated error would be too much.


Top
 Profile  
 
PostPosted: Sun Mar 09, 2014 10:57 pm 
Offline
Member

Joined: Wed Nov 14, 2007 2:12 pm
Posts: 14045
Location: Isleham, Cambridgeshire, UK.
phill.butte wrote:
I just ran a comparison on the two methods and here's what I found for the test pano I did yesterday. The pano was 10 columns by 5 rows with a Canon 60D and 200 mm lens. The image FOV was 5 degrees. The first problem is that the horizontal and vertical overlap figures for the Epic 100 are different for almost all FOV settings, sometimes by as much as 10%. In this case the vertical overlap is 33.2% and the horizontal 29.9%. Papywizard only allows one setting for overlap and it doesn't allow fractions, so I used 33%.
Papywizard calculated that the pitch head movement was 2.32 degrees and the yaw movement was 3.88 degrees for each image. The actual is 2.67 and 4.17. Papywizard calculated the total pitch movement at 11.6 degrees and the total yaw movement at 38 .8 degrees, vs. 13.35 and 41.67 actual. That's a total difference of 13.1% for the pitch and 7% in the yaw axis.
So I would say that for a small panorama using the Papywizard simulator for the Epic 100 would probably be "close enough". For the large, 300-400 gigapixel panos that I like to do I think the accumulated error would be too much.


Thanks for the comparison.

But the end result of the stitch and reneder is the only thing that really matters.

How do the pano images compare after using your XML and that generated by Papywizard?


Top
 Profile  
 
PostPosted: Sun Mar 09, 2014 11:38 pm 
Offline
Member
User avatar

Joined: Sun Oct 21, 2012 2:26 am
Posts: 41
Location: Paradise, California
I get similar results from both XML files for the small (5x10) test pano I did both before and after stitching and rendering. The stitcher corrected all the images with enough texture in them to create control points and the clouds were indistinct enough that the errors didn't matter. I would say it's "good enough". I'll try the comparison again when the weather clears up and I get a chance to do some more gigapixel panoramas.


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 9 posts ] 

All times are UTC + 1 hour


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group