Hi list,
Is i.rectify only designed for down-looking aerial photography? Or can it be used for highly oblique imagery too, for example photos taken side-looking from a plane window, or from a mountain-top covering a valley?
Thanks,
-k.
Hi list,
Is i.rectify only designed for down-looking aerial photography? Or can it be used for highly oblique imagery too, for example photos taken side-looking from a plane window, or from a mountain-top covering a valley?
Thanks,
-k.
Hi Ken,
I have been working on (and struggling with) this as well and did not succeed yet:
For more info and relevant links see: https://lists.osgeo.org/pipermail/grass-user/2015-August/072881.html
My lessons so far are that one
a) needs to use i.ortho.* modules from GRASS 6.4 for this
b) you have to have fairly good estimates of you camera angle (roll, pitch, yaw) (and preferably use as much as possible down-facing angles)
c) according to the manual you need at least 16 GCPs, as much as possible evenly distributed over your image, and it does not seem to work with less than that at all.
In my case the main problem was c), as I only had a few objects I could use for GCPs in my camera viewshed and all of them were more or less in the background. So, I was missing elements I could recognize in foreground. Therefore I placed colored sheets of paper in the camera viewshed, which I measured with GPS, so I can match camera coordinate system and geographical coordinate system.
Paul Kelly used a different approach for that task (which I unfortunately did not manage to try yet):
http://www.stjohnspoint.co.uk/work/thesis/
Cheers
Stefan
-----Original Message-----
From: grass-user [mailto:grass-user-bounces@lists.osgeo.org] On Behalf Of Ken Mankoff
Sent: 28. april 2016 19:49
To: grass-user <grass-user@lists.osgeo.org>
Subject: [GRASS-user] i.rectify and ground-based photos
Hi list,
Is i.rectify only designed for down-looking aerial photography? Or can it be used for highly oblique imagery too, for example photos taken side-looking from a plane window, or from a mountain-top covering a valley?
Thanks,
-k.
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user
Hi Stefan,
Thanks for the links. From reading the threads and other links below (and links therein) it seems like your summary is correct. Except the Kelly thesis and the Neteler (2005) paper show oblique photos (extreme oblique in the thesis). It is a shame this doesn't seem present in GRASS 7, but it is easy enough to work in 6.4 when necessary.
I'm going to look into this more and will write back if I make progress. I would appreciate the same from you. Feel free to contact me off list if you prefer.
I'm also looking into PRACTISE http://www.geosci-model-dev.net/9/307/2016/ and ImGRAFT http://imgraft.glaciology.net/ The latter doesn't work off-the-shelf in Octave, but perhaps the former will, which would be a nice solution. These two tools work in high-relief terrain, which the Kelly thesis mentions and demonstrates also.
-k.
On 2016-04-29 at 02:49, Blumentrath, Stefan <Stefan.Blumentrath@nina.no> wrote:
Hi Ken,
I have been working on (and struggling with) this as well and did not
succeed yet: For more info and relevant links see:
https://lists.osgeo.org/pipermail/grass-user/2015-August/072881.htmlMy lessons so far are that one a) needs to use i.ortho.* modules from
GRASS 6.4 for this b) you have to have fairly good estimates of you
camera angle (roll, pitch, yaw) (and preferably use as much as
possible down-facing angles) c) according to the manual you need at
least 16 GCPs, as much as possible evenly distributed over your image,
and it does not seem to work with less than that at all.In my case the main problem was c), as I only had a few objects I
could use for GCPs in my camera viewshed and all of them were more or
less in the background. So, I was missing elements I could recognize
in foreground. Therefore I placed colored sheets of paper in the
camera viewshed, which I measured with GPS, so I can match camera
coordinate system and geographical coordinate system.Paul Kelly used a different approach for that task (which I
unfortunately did not manage to try yet):
http://www.stjohnspoint.co.uk/work/thesis/-----Original Message----- From: grass-user
[mailto:grass-user-bounces@lists.osgeo.org] On Behalf Of Ken Mankoff
Sent: 28. april 2016 19:49 To: grass-user <grass-user@lists.osgeo.org>
Subject: [GRASS-user] i.rectify and ground-based photosHi list,
Is i.rectify only designed for down-looking aerial photography? Or can
it be used for highly oblique imagery too, for example photos taken
side-looking from a plane window, or from a mountain-top covering a
valley?Thanks,
-k.