Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.os.linux.misc > #13129 > unrolled thread
| Started by | Eli the Bearded <*@eli.users.panix.com> |
|---|---|
| First post | 2014-12-30 00:45 +0000 |
| Last post | 2015-01-04 00:58 +0000 |
| Articles | 6 — 2 participants |
Back to article view | Back to comp.os.linux.misc
still image stitched from video? Eli the Bearded <*@eli.users.panix.com> - 2014-12-30 00:45 +0000
Re: still image stitched from video? Jack Strangio <jackstrangio@yahoo.com> - 2014-12-31 02:18 +0000
Re: still image stitched from video? Eli the Bearded <*@eli.users.panix.com> - 2014-12-31 21:57 +0000
Re: still image stitched from video? Jack Strangio <jackstrangio@yahoo.com> - 2015-01-01 03:23 +0000
Re: still image stitched from video? Eli the Bearded <*@eli.users.panix.com> - 2015-01-02 22:58 +0000
Re: still image stitched from video? Jack Strangio <jackstrangio@yahoo.com> - 2015-01-04 00:58 +0000
| From | Eli the Bearded <*@eli.users.panix.com> |
|---|---|
| Date | 2014-12-30 00:45 +0000 |
| Subject | still image stitched from video? |
| Message-ID | <eli$1412291932@qz.little-neck.ny.us> |
Let's say you have a clip of video that pans over a landscape. What would be the easiest way to make a "photo" that consists of the stitched frames? Is this something Hugin can do? I've had very poor results with Hugin in the past, but haven't tried it in a couple of years. The image links from this don't work, but the problem is described: https://www.flickr.com/groups/hugin/discuss/72157630021562085/ I never got that mosaic to work. The photos were of the chess mural on Plum St, in San Francisco. A street view link that may or may not work: https://www.google.com/maps/@37.7709283,-122.4183732,3a,75y,180.35h,83.64t/data=!3m4!1e1!3m2!1sJDALROS8-kvrJWPa-qN9fQ!2e0 Elijah ------ has some Yosemite video in mind for this new image stitch project
[toc] | [next] | [standalone]
| From | Jack Strangio <jackstrangio@yahoo.com> |
|---|---|
| Date | 2014-12-31 02:18 +0000 |
| Message-ID | <nHFD2C.2LE@yahoo.com> |
| In reply to | #13129 |
Eli the Bearded <*@eli.users.panix.com> writes:
> Let's say you have a clip of video that pans over a landscape. What
> would be the easiest way to make a "photo" that consists of the stitched
> frames?
>
You should be able to extract the frames with something like
mplayer -vo jpeg <video filename>
then select every 10th (or whatever) frame, and stitch those together.
convert *.jpg file.pdf
or something similar.
Jack
--
"I'm a home-loving girl. And that's where I wish I was."
"At home ..."
"Loving."
- Laugh-In, 1968
[toc] | [prev] | [next] | [standalone]
| From | Eli the Bearded <*@eli.users.panix.com> |
|---|---|
| Date | 2014-12-31 21:57 +0000 |
| Message-ID | <eli$1412311657@qz.little-neck.ny.us> |
| In reply to | #13137 |
In comp.os.linux.misc, Jack Strangio <jackstrangio@yahoo.com> wrote: > Eli the Bearded <*@eli.users.panix.com> writes: >> Let's say you have a clip of video that pans over a landscape. What >> would be the easiest way to make a "photo" that consists of the stitched >> frames? > You should be able to extract the frames with something like > mplayer -vo jpeg <video filename> Yeah, I'm intimately familiar with how to use mplayer for that purpose. Generally my command line is closer to: mplayer -ss TIME -endpos TIME -nosound -vo jpeg:outdir=DIRECTORY VIDEO > then select every 10th (or whatever) frame, and stitch those together. > convert *.jpg file.pdf That's not "stitch"ing. That creates a PDF file with one image per page. Stitching actually combines multiple images into a single unified image. When I extract JPEGs out of the video files my camera produces, all of the lens metadata is lost. $ exiftool /mnt/dcim/860_1226/mvi_8788.mov [194 lines of output including :] Aperture Value : 9.0 Focal Length : 11.2 mm Lens : 5.0 - 25.0 mm Lens ID : Unknown 5-25mm Rotation : 0 Video Frame Rate : 29.97 $ mplayer -endpos 5 -vo jpeg:outdir=/tmp/demo $ exiftool /tmp/demo/00000100.jpg [ 23 lines output with no lens/focal/aperture info ] From experience, I know that stitching programs *really* want lens and focal length to do the math to correct for distortion between frames. That's a big reason why I'd like like to be able to use video files directly. That video includes both underwater and above water segments would would provide further stitching problems, since the index of refraction changes mid-way. But, hey, I'd like to see what a program could do with it. Elijah ------ loves his waterproof "rugged" camera
[toc] | [prev] | [next] | [standalone]
| From | Jack Strangio <jackstrangio@yahoo.com> |
|---|---|
| Date | 2015-01-01 03:23 +0000 |
| Message-ID | <nHHArs.Jys@yahoo.com> |
| In reply to | #13144 |
Eli the Bearded <*@eli.users.panix.com> writes:
> In comp.os.linux.misc, Jack Strangio <jackstrangio@yahoo.com> wrote:
> > Eli the Bearded <*@eli.users.panix.com> writes:
> >> Let's say you have a clip of video that pans over a landscape. What
> >> would be the easiest way to make a "photo" that consists of the stitched
> >> frames?
> > You should be able to extract the frames with something like
> > mplayer -vo jpeg <video filename>
>
> Yeah, I'm intimately familiar with how to use mplayer for that purpose.
> Generally my command line is closer to:
>
> mplayer -ss TIME -endpos TIME -nosound -vo jpeg:outdir=DIRECTORY VIDEO
>
>
> > then select every 10th (or whatever) frame, and stitch those together.
> > convert *.jpg file.pdf
>
> That's not "stitch"ing. That creates a PDF file with one image per page.
> Stitching actually combines multiple images into a single unified image.
>
> When I extract JPEGs out of the video files my camera produces, all of
> the lens metadata is lost.
Very true. I was being lazy with my memory. hugin would do a better job
and it allows some parameters to be played with.
http://hugin.sourceforge.net/tutorials/index.shtml
hugin should be available by selecting it in your package-manager.
Jack
--
"I'm a home-loving girl. And that's where I wish I was."
"At home ..."
"Loving."
- Laugh-In, 1968
[toc] | [prev] | [next] | [standalone]
| From | Eli the Bearded <*@eli.users.panix.com> |
|---|---|
| Date | 2015-01-02 22:58 +0000 |
| Message-ID | <eli$1501021758@qz.little-neck.ny.us> |
| In reply to | #13151 |
In comp.os.linux.misc, Jack Strangio <jackstrangio@yahoo.com> wrote: > Eli the Bearded <*@eli.users.panix.com> writes: > > When I extract JPEGs out of the video files my camera produces, all of > > the lens metadata is lost. > Very true. I was being lazy with my memory. hugin would do a better job > and it allows some parameters to be played with. Does Hugin do this from video? > http://hugin.sourceforge.net/tutorials/index.shtml A quick search of the documentation does not answer the question. Elijah ------ the hugin release notes are also pretty terse
[toc] | [prev] | [next] | [standalone]
| From | Jack Strangio <jackstrangio@yahoo.com> |
|---|---|
| Date | 2015-01-04 00:58 +0000 |
| Message-ID | <nHMo1H.9KB@yahoo.com> |
| In reply to | #13186 |
Eli the Bearded <*@eli.users.panix.com> writes:
> In comp.os.linux.misc, Jack Strangio <jackstrangio@yahoo.com> wrote:
> > and it allows some parameters to be played with.
>
> Does Hugin do this from video?
>
That's more sophisticated than my level of expertise. Does this help?
http://hugin.sourceforge.net/docs/manual/Lens_correction_model.html
Regards,
Jack
--
"I'm a home-loving girl. And that's where I wish I was."
"At home ..."
"Loving."
- Laugh-In, 1968
[toc] | [prev] | [standalone]
Back to top | Article view | comp.os.linux.misc
csiph-web