Re: [hugin-ptx] No succes stitch spherical pano from Meike 6.5 mm circular fisheye.

[Greg 'groggy' Lehey]

On Friday,  2 October 2020 at  6:50:35 -0700, Henk Tijdink wrote:
> I have send the pictures with Wetransfer.

OK, I have good news and not-so-good news.

First the good news: the panos stitched out of the box with no
problems.  First I had to tell Hugin about the lens: circular fisheye,
190° FoV.  You can see the results at
http://www.lemis.com/grog/photos/Photos.php?dirdate=20201002.

The not-so-good news: the fast panorama preview makes a mess of the
display, and may be the reason that you thought it didn't work.  But
the normal preview (the one nobody uses) shows the display correctly.
This is probably a bug that needs to be corrected.

Greg
--
Sent from my desktop computer.
Finger groog...@gmail.com for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed.  If your Microsoft mail program
reports problems, please read http://lemis.com/broken-MUA

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/20201003013805.GA71763%40eureka.lemis.com.


signature.asc
Description: PGP signature

[hugin-ptx] Re: No succes stitch spherical pano from Meike 6.5 mm circular fisheye.

[T. Modes]

Hi Henk,

Am Donnerstag, 1. Oktober 2020 16:33:02 UTC+2 schrieb Henk Tijdink:
>
>
> In PTGUI is the lens in te database.
> It gives the following values for the Meike lens. This for the Jpeg files 
> from the camera
> F  theoretic  6,5 mm After optimizing are the 
> lens parameters
> Practical  6,66 mm   a= -0,093   and 
> sensor shift.
> Fov degrees 181,01   b=  0,18  d= 
> -0,18%
> Fisheye factor-0,824152 c= -0,11   e= 
> -0,17%
>
 

>
> Is it possible to "translate" the lens parameters of this from the PTGUI 
> database to hugin? 
>

This is not so easy. PTGUI fisheye factor does not translate 1:1 to a 
projection of libpano. The fisheye factor of -0.824 would result in 
something between orthographic and equisolid projection (see 
https://wiki.panotools.org/Fisheye_Projection).
The next pitfall is the fov: PTGUI calculates the fov to the cropped 
circle, libpano/Hugin is using the horizontal dimension of the full image 
(see https://wiki.panotools.org/Panotools_internals#Cropping ). 

I had a look at your pto file. If I remove the control points in the nadir 
area it optimizes nicely. Without the images it is difficult to give 
further advices, but I would check the control points in the nadir area 
(maybe some repetitive pattern has confused the control point finder) and 
then try again.

Thomas

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/37ce77a9-40f9-4f2e-b10b-0d988dd88271o%40googlegroups.com.

[hugin-ptx] Re: Auto calculation of Yaw, pitch and roll failed in some computers

[T. Modes]

Hi Parasar,

Am Freitag, 2. Oktober 2020 06:46:17 UTC+2 schrieb Parasar Ghimire:
>
> Hi Thomas,
> Thanks for replying. Please find the attached images. This set of pictures 
> works on my computer when I align them but doesn't work on other computers. 
> Please advise me if I am missing something.
> My general approach is:
> 1. import images into HUGIN
> 2. Set Lens type as Circular fisheye and focal length as 3.154
> 3. Click Align to auto-detect control points. 
>

The main problem with the provided sample set is that some images have been 
rotated physically and some not (images 1-6 are portrait, images 8 and 9 
landscape).
This results that when you import the images into Hugin 2 different lenses 
are assigned and when you set the lens type to circular fisheye only the 
first lens is changed.
This confuses the control point finder and it fails to align.

If I manually rotate the last 2 images also to portrait format, the 
assistant works fine and aligns the pano. There is enough overlap for Hugin

If this does not work on all computers then on these computer the 
camera/lens is already in the local camera/lens database with different 
settings and when importing the images these settings are read from the 
database and used then. 
So after importing the images check the lens projection, focal length and 
multiplier on these computer after importing images.

Thomas

@Luis: Hugin is not using the lensfun database. So this hint does not apply.

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/7676bf9b-bfbd-424f-a588-6234c9a24ba7o%40googlegroups.com.

Re: [hugin-ptx] Re: Auto calculation of Yaw, pitch and roll failed in some computers

[Luís Henrique Camargo Quiroz]

   Maybe discrepancies in lensfun? Not all the same version, not all
updated?

  Run lensfun-update-data.  Hugin only can benefit from good lens
parameters if they can be found in the database.
http://lensfun.sourceforge.net/usage/

  Luís Henrique

Em sex., 2 de out. de 2020 às 07:38, giuseppe porciani <
giuseppe.porci...@gmail.com> escreveu:

> Hi,
> I gave it a try. In my case Hugin does not ask for lens parameters because
> the exif data give rectilinear: 98 degrees Hfov, which seems the case
> considering the relatively low distortion. In the five equatorial pictures
> CP detects enough control points. It does not in the zenith and nadir
> pictures. Adding manually control points in these pictures should work.
> Perhaps problems come from the not correct parameters given for the lens. ..
> Regards,
>
> Giuseppe
>
> Il giorno ven 2 ott 2020 alle ore 08:48 Greg 'groggy' Lehey <
> groog...@gmail.com> ha scritto:
>
>> On Thursday,  1 October 2020 at 21:46:17 -0700, Parasar Ghimire wrote:
>> > Hi Thomas,
>> > Thanks for replying. Please find the attached images. This set of
>> pictures
>> > works on my computer when I align them but doesn't work on other
>> computers.
>> > Please advise me if I am missing something.
>> > My general approach is:
>> > 1. import images into HUGIN
>> > 2. Set Lens type as Circular fisheye and focal length as 3.154
>> > 3. Click Align to auto-detect control points.
>> >
>> > The results are fine on my computers but fail in others as it isn't
>> able to
>> > calculate yaw, pitch, and roll values. This is why I cannot recommend
>> and
>> > standardize the process.
>> >
>> >
>> https://drive.google.com/file/d/1AlkHTbkZIKkAhTS1QxT4Uh5SQLl6KV12/view?usp=sharing
>>
>> OK, I've taken a look at the images.  I'm amazed that you can get them
>> to align at all.  There's barely any overlap.  And most certainly the
>> images as shown were not taken with a circular fisheye.  I selected
>> full frame fisheye (which is probably also stretching the case) and
>> got the first five without trouble.  I needed to add the sixth
>> manually.  And that's as far as I got.  I couldn't recognize any
>> connection between the last two and the rest.  I think it speaks
>> highly for Hugin that any version works.
>>
>> Why don't you try again with more images?  Consider that the control
>> point detector needs to find at least 6 control points between images.
>> Think what points might be of use and ensure that they appear in at
>> least two images.  If all else fails (like the ground in your images)
>> you could consider putting something like a lens cap on the ground and
>> taking it in different images, then removing the cap and taking more
>> images, finally omitting the images with the lens cap from the final
>> panorama.
>>
>> And why does this work on some computers and not others?  I've seen
>> similar issues in the past, but first you should check if you have the
>> same control point detectors on all the computers.
>>
>> Greg
>> --
>> Sent from my desktop computer.
>> Finger groog...@gmail.com for PGP public key.
>> See complete headers for address and phone numbers.
>> This message is digitally signed.  If your Microsoft mail program
>> reports problems, please read http://lemis.com/broken-MUA
>>
>> --
>> A list of frequently asked questions is available at:
>> http://wiki.panotools.org/Hugin_FAQ
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "hugin and other free panoramic software" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to hugin-ptx+unsubscr...@googlegroups.com.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/hugin-ptx/20201002064845.GB2619%40eureka.lemis.com
>> .
>>
> --
> A list of frequently asked questions is available at:
> http://wiki.panotools.org/Hugin_FAQ
> ---
> You received this message because you are subscribed to the Google Groups
> "hugin and other free panoramic software" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to hugin-ptx+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/hugin-ptx/CAO0N8z0qtO9-_FVC1t5mOK6oSSAaCU%3DGDRqNHfKr360%2B9uNVNw%40mail.gmail.com
> 
> .
>


-- 
-- 
Luis Henrique Camargo Quiroz
http://luishcq.br.tripod.com - http://www.christusrex.org/www2/cantgreg
http://panoramaslh.net/

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit 

Re: [hugin-ptx] Re: Auto calculation of Yaw, pitch and roll failed in some computers

[giuseppe porciani]

Hi,
Sorry, I made some mistakes and bad assumptions.
I tried with these lens parameters: fisheye full frame; focal length 2,8;
crop factor 5,6.
It stitches. From the exif data it seems you use a gopro 6 black. I took
sensor and lens data from specs of this gopro version. Hope that may help.
Regards,

Giuseppe

Il giorno ven 2 ott 2020 alle ore 12:38 giuseppe porciani <
giuseppe.porci...@gmail.com> ha scritto:

> Hi,
> I gave it a try. In my case Hugin does not ask for lens parameters because
> the exif data give rectilinear: 98 degrees Hfov, which seems the case
> considering the relatively low distortion. In the five equatorial pictures
> CP detects enough control points. It does not in the zenith and nadir
> pictures. Adding manually control points in these pictures should work.
> Perhaps problems come from the not correct parameters given for the lens. ..
> Regards,
>
> Giuseppe
>
> Il giorno ven 2 ott 2020 alle ore 08:48 Greg 'groggy' Lehey <
> groog...@gmail.com> ha scritto:
>
>> On Thursday,  1 October 2020 at 21:46:17 -0700, Parasar Ghimire wrote:
>> > Hi Thomas,
>> > Thanks for replying. Please find the attached images. This set of
>> pictures
>> > works on my computer when I align them but doesn't work on other
>> computers.
>> > Please advise me if I am missing something.
>> > My general approach is:
>> > 1. import images into HUGIN
>> > 2. Set Lens type as Circular fisheye and focal length as 3.154
>> > 3. Click Align to auto-detect control points.
>> >
>> > The results are fine on my computers but fail in others as it isn't
>> able to
>> > calculate yaw, pitch, and roll values. This is why I cannot recommend
>> and
>> > standardize the process.
>> >
>> >
>> https://drive.google.com/file/d/1AlkHTbkZIKkAhTS1QxT4Uh5SQLl6KV12/view?usp=sharing
>>
>> OK, I've taken a look at the images.  I'm amazed that you can get them
>> to align at all.  There's barely any overlap.  And most certainly the
>> images as shown were not taken with a circular fisheye.  I selected
>> full frame fisheye (which is probably also stretching the case) and
>> got the first five without trouble.  I needed to add the sixth
>> manually.  And that's as far as I got.  I couldn't recognize any
>> connection between the last two and the rest.  I think it speaks
>> highly for Hugin that any version works.
>>
>> Why don't you try again with more images?  Consider that the control
>> point detector needs to find at least 6 control points between images.
>> Think what points might be of use and ensure that they appear in at
>> least two images.  If all else fails (like the ground in your images)
>> you could consider putting something like a lens cap on the ground and
>> taking it in different images, then removing the cap and taking more
>> images, finally omitting the images with the lens cap from the final
>> panorama.
>>
>> And why does this work on some computers and not others?  I've seen
>> similar issues in the past, but first you should check if you have the
>> same control point detectors on all the computers.
>>
>> Greg
>> --
>> Sent from my desktop computer.
>> Finger groog...@gmail.com for PGP public key.
>> See complete headers for address and phone numbers.
>> This message is digitally signed.  If your Microsoft mail program
>> reports problems, please read http://lemis.com/broken-MUA
>>
>> --
>> A list of frequently asked questions is available at:
>> http://wiki.panotools.org/Hugin_FAQ
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "hugin and other free panoramic software" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to hugin-ptx+unsubscr...@googlegroups.com.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/hugin-ptx/20201002064845.GB2619%40eureka.lemis.com
>> .
>>
>

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/CAO0N8z05F7EoYd8DYYH4RgEqu54HYrkuAaPCCzsQSgPZXbQ6CA%40mail.gmail.com.

Re: [hugin-ptx] Re: Auto calculation of Yaw, pitch and roll failed in some computers

[giuseppe porciani]

Hi,
I gave it a try. In my case Hugin does not ask for lens parameters because
the exif data give rectilinear: 98 degrees Hfov, which seems the case
considering the relatively low distortion. In the five equatorial pictures
CP detects enough control points. It does not in the zenith and nadir
pictures. Adding manually control points in these pictures should work.
Perhaps problems come from the not correct parameters given for the lens. ..
Regards,

Giuseppe

Il giorno ven 2 ott 2020 alle ore 08:48 Greg 'groggy' Lehey <
groog...@gmail.com> ha scritto:

> On Thursday,  1 October 2020 at 21:46:17 -0700, Parasar Ghimire wrote:
> > Hi Thomas,
> > Thanks for replying. Please find the attached images. This set of
> pictures
> > works on my computer when I align them but doesn't work on other
> computers.
> > Please advise me if I am missing something.
> > My general approach is:
> > 1. import images into HUGIN
> > 2. Set Lens type as Circular fisheye and focal length as 3.154
> > 3. Click Align to auto-detect control points.
> >
> > The results are fine on my computers but fail in others as it isn't able
> to
> > calculate yaw, pitch, and roll values. This is why I cannot recommend and
> > standardize the process.
> >
> >
> https://drive.google.com/file/d/1AlkHTbkZIKkAhTS1QxT4Uh5SQLl6KV12/view?usp=sharing
>
> OK, I've taken a look at the images.  I'm amazed that you can get them
> to align at all.  There's barely any overlap.  And most certainly the
> images as shown were not taken with a circular fisheye.  I selected
> full frame fisheye (which is probably also stretching the case) and
> got the first five without trouble.  I needed to add the sixth
> manually.  And that's as far as I got.  I couldn't recognize any
> connection between the last two and the rest.  I think it speaks
> highly for Hugin that any version works.
>
> Why don't you try again with more images?  Consider that the control
> point detector needs to find at least 6 control points between images.
> Think what points might be of use and ensure that they appear in at
> least two images.  If all else fails (like the ground in your images)
> you could consider putting something like a lens cap on the ground and
> taking it in different images, then removing the cap and taking more
> images, finally omitting the images with the lens cap from the final
> panorama.
>
> And why does this work on some computers and not others?  I've seen
> similar issues in the past, but first you should check if you have the
> same control point detectors on all the computers.
>
> Greg
> --
> Sent from my desktop computer.
> Finger groog...@gmail.com for PGP public key.
> See complete headers for address and phone numbers.
> This message is digitally signed.  If your Microsoft mail program
> reports problems, please read http://lemis.com/broken-MUA
>
> --
> A list of frequently asked questions is available at:
> http://wiki.panotools.org/Hugin_FAQ
> ---
> You received this message because you are subscribed to the Google Groups
> "hugin and other free panoramic software" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to hugin-ptx+unsubscr...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/hugin-ptx/20201002064845.GB2619%40eureka.lemis.com
> .
>

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/CAO0N8z0qtO9-_FVC1t5mOK6oSSAaCU%3DGDRqNHfKr360%2B9uNVNw%40mail.gmail.com.

Re: [hugin-ptx] Re: Auto calculation of Yaw, pitch and roll failed in some computers

[Greg 'groggy' Lehey]

On Thursday,  1 October 2020 at 21:46:17 -0700, Parasar Ghimire wrote:
> Hi Thomas,
> Thanks for replying. Please find the attached images. This set of pictures
> works on my computer when I align them but doesn't work on other computers.
> Please advise me if I am missing something.
> My general approach is:
> 1. import images into HUGIN
> 2. Set Lens type as Circular fisheye and focal length as 3.154
> 3. Click Align to auto-detect control points.
>
> The results are fine on my computers but fail in others as it isn't able to
> calculate yaw, pitch, and roll values. This is why I cannot recommend and
> standardize the process.
>
> https://drive.google.com/file/d/1AlkHTbkZIKkAhTS1QxT4Uh5SQLl6KV12/view?usp=sharing

OK, I've taken a look at the images.  I'm amazed that you can get them
to align at all.  There's barely any overlap.  And most certainly the
images as shown were not taken with a circular fisheye.  I selected
full frame fisheye (which is probably also stretching the case) and
got the first five without trouble.  I needed to add the sixth
manually.  And that's as far as I got.  I couldn't recognize any
connection between the last two and the rest.  I think it speaks
highly for Hugin that any version works.

Why don't you try again with more images?  Consider that the control
point detector needs to find at least 6 control points between images.
Think what points might be of use and ensure that they appear in at
least two images.  If all else fails (like the ground in your images)
you could consider putting something like a lens cap on the ground and
taking it in different images, then removing the cap and taking more
images, finally omitting the images with the lens cap from the final
panorama.

And why does this work on some computers and not others?  I've seen
similar issues in the past, but first you should check if you have the
same control point detectors on all the computers.

Greg
--
Sent from my desktop computer.
Finger groog...@gmail.com for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed.  If your Microsoft mail program
reports problems, please read http://lemis.com/broken-MUA

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/20201002064845.GB2619%40eureka.lemis.com.


signature.asc
Description: PGP signature

Re: [hugin-ptx] No succes stitch spherical pano from Meike 6.5 mm circular fisheye.

[Greg 'groggy' Lehey]

On Thursday,  1 October 2020 at  7:33:02 -0700, Henk Tijdink wrote:
> Dear readers,
> I've bought a Meike circular fisheyelens 6.5 mm f2 for a MFT camera. and
> want to use it for making full spherical panorama's.
> You get than partly a circle  On the long side of the sensor it is shaved
> off. FOV there ca 120 degrees.
> Making 4 pictures at 90 degrees turn and in the demo of PTGUI it stitches
> perfectly.
> In Hugin however I can't get it all right.
> It finds control points between the images. I did the circular crop in the
> mask tab, but optimizing gives a very bad result.

I do this kind of panorama all the time using an Olympus 8 mm f/1.8
fisheye, also with 4 images, and it works fine.  If you can make the
images available online (NOT on this list!), I'll take a look at them,
at least in part because I was rather interested in this lens.

Greg
--
Sent from my desktop computer.
Finger groog...@gmail.com for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed.  If your Microsoft mail program
reports problems, please read http://lemis.com/broken-MUA

-- 
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
--- 
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/hugin-ptx/20201002061210.GA2619%40eureka.lemis.com.


signature.asc
Description: PGP signature