Work in progress

This is a case study to match the Hugin workflow page.

pto_gen, from command line

The first step reads the source files sufficiently to generate a project file:

The error messages make no sense, and are probably a harmless bug.

After running pto_gen, the project file contains:

A number of commented-out parameters at the end have been omitted.

The meaning of this file is:

p (panorama) line

p f2 w3000 h1500 v360 E11.9366 R0 n"TIFF_m c:LZW r:CROP"

• f2: Projection 2, default: Equirectangular
• w3000: Width 3000 pixels
• h1500: Height 1500 pixels
• v360: Horizontal field of view 360° (default)
• E11.9366: Exposure value for final panorama. Where does this come from?
• R0: Stitching mode 0 (normal LDR mode).
• n"TIFF_m c:LZW r:CROP": File format (TIFF, multi-file, one image per file), LZW compression, cropped output.

m (mode) line

• g1: Gamma for internal computations (1.0, default).
• i0: Interpolator (poly3, default).
• f0: Not defined
• m2: Not defined
• p0.00784314: Not defined

i (image) lines

These are, not surprisingly, all similar. The first is almost enough:

Here's the Exif data for this image:

The fields of view are incorrect, what you would expect of a rectilinear lens of that focal length. In fact, the horizontal field of view is 123.9°. They're calculated in this output, so not of importance.

• w5182: Image width in pixels
• h3886: Image height in pixels
• f3: Projection (full frame fisheye). This is interesting because pto_gen recognizes it, but in the GUI it needs to be specified.
• v94.4078430468941: Horizontal field of view (degrees). Again, it has been calculated for a rectilinear lens.
• Ra0: EMoR photometric model parameter, not documented, but default value.
• Rb0: EMoR photometric model parameter, not documented, but default value.
• Rc0: EMoR photometric model parameter, not documented, but default value.
• Rd0: EMoR photometric model parameter, not documented, but default value.
• Re0: EMoR photometric model parameter, not documented, but default value.
• Eev11.9366378213403: Image exposure in EV. It's not clear where this comes from. It's just shy of 1 EV less than the information in the Exif data.
• Er1: White balance factor for red channel.
• Eb1: White balance factor for blue channel.
• r0: roll angle
• p0: pitch angle
• y0: yaw angle
• TrX0: Mosaic mode translation offset, presumably X axis.
• TrY0: Mosaic mode translation offset, presumably Y axis.
• TrZ0: Mosaic mode translation offset, presumably Z axis.
• Tpy0: orientation of remapping plane for translation offsets (which?)
• Tpp0: orientation of remapping plane for translation offsets (which?)
• j0: Stack number.
• a0: lens correction coefficient, not used here.
• b0: lens correction coefficient, not used here.
• c0: lens correction coefficient, not used here.
• d0: initial horizontal lens offset, not used here.
• e0: initial vertical lens offset, not used here.
• g0: initial horizontal lens shear, not used here.
• t0: initial vertical lens shear, not used here.
• Va1: vignetting correction coefficient, not used here.
• Vb0: vignetting correction coefficient, not used here.
• Vc0: vignetting correction coefficient, not used here.
• Vd0: vignetting correction coefficient, not used here.
• Vx0: radial horizontal vignetting correction offset, not used here.
• Vy0: radial vertical vignetting correction offset, not used here.
• Vm5: vignetting correction mode (radial polynomial correction by division)
• n"e-from-house-0.tiff": Input file name

Somewhere here I'm missing lens information. I could have sworn I saw it before, but it's not in nona.txt and it's not in the project file.

Read images into GUI

Reading the images into the GUI does a little more than just run pto_gen. In fact, running ktrace suggests that it's all done internally. The result of reading in the files and saving a project file shows some differences (e-from-house-0-e-from-house-3-GUI.pto):

p line

Here the horizontal and vertical sizes have been reduced, by different factors. Why?

i lines

Here the y parameter (yaw) has been set. The values for the four images are -106.208823427756, 35.4029411425853, 35.4029411425853 and 106.208823427756. That's not the correct relationship (they were taken at steps of 90°), so maybe it relates to where they appeared on the screen.

v lines

The big difference is in the v lines, which were not present in the pto_gen output:

I haven't found documentation for these yet.

Running cpfind

Describe CPs.

No wraparound.

PTO files: -rw-r--r-- 1 grog lemis 9,350 18 Jun 17:50 e-from-house-0-e-from-house-3-GUI-cpfound.pto -rw-r--r-- 1 grog wheel 2,121 18 Jun 17:08 e-from-house-0-e-from-house-3-GUI.pto -rwxr--r-- 1 grog lemis 9,336 18 Jun 18:02 e-from-house-0-e-from-house-3-Microsoft-cpfound.pto -rw-r--r-- 1 grog lemis 1,978 17 Jun 14:25 e-from-house-0-e-from-house-3.pto pto_gen 19 June Run GUI, cpfind: === grog@eureka (/dev/pts/28) /photowork/Hugin-build-eureka 92 -> ps aux|grep cpfind grog 69358 49.9 2.0 801448 671604 - RNs 12:22pm 0:06.34 cpfind --multirow -o /tmp/ap_res6w6ICm /tmp/ap_inprojv8AcxR grog 69360 0.0 0.0 18848 2388 28 S+ 12:22pm 0:00.00 grep cpfind Run manually: === grog@eureka (/dev/pts/28) /photowork/Hugin-build-eureka 95 -> cpfind --multirow -o /tmp/ap_res6w6ICm /tmp/ap_inprojv8AcxR Hugin's cpfind 2018.0.0.5abfb4de7961 based on Pan-o-matic by Anael Orlinski Project contains the following images: Image 0 Imagefile: /photowork/Hugin-build-eureka/00 Remapped : yes Image 1 Imagefile: /photowork/Hugin-build-eureka/01 Remapped : yes Image 2 Imagefile: /photowork/Hugin-build-eureka/02 Remapped : yes Image 3 Imagefile: /photowork/Hugin-build-eureka/03 Remapped : yes --- Analyze Images --- i0 : Analyzing image... Warning: no TIFFTAG_SAMPLEFORMAT or TIFFTAG_DATATYPE, guessing pixeltype 'UINT16'. Warning: no TIFFTAG_SAMPLEFORMAT or TIFFTAG_DATATYPE, guessing pixeltype 'UINT16'. i1 : Analyzing image... Warning: no TIFFTAG_SAMPLEFORMAT or TIFFTAG_DATATYPE, guessing pixeltype 'UINT16'. Warning: no TIFFTAG_SAMPLEFORMAT or TIFFTAG_DATATYPE, guessing pixeltype 'UINT16'. i2 : Analyzing image... Warning: no TIFFTAG_SAMPLEFORMAT or TIFFTAG_DATATYPE, guessing pixeltype 'UINT16'. Warning: no TIFFTAG_SAMPLEFORMAT or TIFFTAG_DATATYPE, guessing pixeltype 'UINT16'. i3 : Analyzing image... Warning: no TIFFTAG_SAMPLEFORMAT or TIFFTAG_DATATYPE, guessing pixeltype 'UINT16'. Warning: no TIFFTAG_SAMPLEFORMAT or TIFFTAG_DATATYPE, guessing pixeltype 'UINT16'. --- Find matches --- i0 <> i1 : Found 20 matches i1 <> i2 : Found 23 matches i2 <> i3 : Found 19 matches --- Find matches for overlapping images --- i0 <> i2 : Found 0 matches i1 <> i3 : Found 0 matches --- Write Project output --- Written output to /tmp/ap_res6w6ICm Detection took 23.404 seconds. --multirow

---

Greg's home page

Greg's diary

Greg's photos

Copyright

$Id: skel.php,v 1.7 2014/02/16 02:49:28 grog Exp $