Image processing project - getting started

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Image processing project - getting started

Justin Close
Hello all.  I am new to the list and I am looking for some guidance and
pointers.

I am a computer science student working on a project involving
manipulation of a large number of images.  The images are aerial shots
of the cities at night (taken from space, actually).  We are taking the
RAW camera images, saving to TIFF, using the date/time stamp of the
image to generally locate what/where the image is of, then manually
morphing the image to scale/rotate/position it accurately to reality.
Reality in this case is Google Earth; the images will eventually be
served up as Google Earth layers.

The CS part of this is how to automate this process.  We wanted to embed
the lat/long information we got back into the image, and GeoTIFF seems
to be the perfect format for this.  Low and behold, it can also contain
morphing information too, it seems.  And there are all these nifty tools
out there that can do this manipulation, apparently.  But for me, who is
just starting out with it all, it seems a bit overwhelming; I don't know
where to start in order to find out what tools I should be using.
Libgeotiff seems like it has a bunch of things that we (it is a team
project) should be using; we don't want to reinvent the wheel (not
completely anyway - a bit of learning by doing can be good, though).

So far the images have just been processed within GE itself, using its
Image overlay features.  But that is tedious and awkward.  First step is
perhaps a better manual tool that can achieve the same results (an image
that is rotated and scaled to fit GE accurately).  A next step is to try
and get those manual operations automated so that a large number of
images can be done quickly, without intervention.

For the tagging (approximate lat/long) I was thinking of something to
read the TIFF tag data, interface with the database that will yield the
lat/long of the image, and then writing that back to the TIFF (now that
I know something about GeoTIFF, probably as a GeoTIFF now instead of
TIFF).  It looks like listgeo/geotifcp are the tools I would want to be
using for doing that.

So, I am just asking for some guidance from folks who have used these
tools and could comment on what might work well for what I am looking
for.  If you want more details, let me know.

Thanks,
Justin




_______________________________________________
Geotiff mailing list
[hidden email]
http://lists.maptools.org/mailman/listinfo/geotiff

signature.asc (563 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Image processing project - getting started

Kyle Shannon
Justin,
I would look at GDAL, gdal.org.  There are a few things that may be of use for you.  There are several command line utilities that may be of use and good a good api for C/C++, python and other languages.  The documentation is good too.  Read through the website and see if it works for you.

kss

# ============================
Kyle Shannon
Physical Science Technician
RMRS Fire Sciences Lab
Fire, Fuels & Smoke - RWU 4405
5775 Highway 10 W.
Missoula, MT 59808
(406)829-6954
[hidden email]
# ============================


On Mon, Nov 8, 2010 at 10:32 AM, Justin Close <[hidden email]> wrote:
Hello all.  I am new to the list and I am looking for some guidance and
pointers.

I am a computer science student working on a project involving
manipulation of a large number of images.  The images are aerial shots
of the cities at night (taken from space, actually).  We are taking the
RAW camera images, saving to TIFF, using the date/time stamp of the
image to generally locate what/where the image is of, then manually
morphing the image to scale/rotate/position it accurately to reality.
Reality in this case is Google Earth; the images will eventually be
served up as Google Earth layers.

The CS part of this is how to automate this process.  We wanted to embed
the lat/long information we got back into the image, and GeoTIFF seems
to be the perfect format for this.  Low and behold, it can also contain
morphing information too, it seems.  And there are all these nifty tools
out there that can do this manipulation, apparently.  But for me, who is
just starting out with it all, it seems a bit overwhelming; I don't know
where to start in order to find out what tools I should be using.
Libgeotiff seems like it has a bunch of things that we (it is a team
project) should be using; we don't want to reinvent the wheel (not
completely anyway - a bit of learning by doing can be good, though).

So far the images have just been processed within GE itself, using its
Image overlay features.  But that is tedious and awkward.  First step is
perhaps a better manual tool that can achieve the same results (an image
that is rotated and scaled to fit GE accurately).  A next step is to try
and get those manual operations automated so that a large number of
images can be done quickly, without intervention.

For the tagging (approximate lat/long) I was thinking of something to
read the TIFF tag data, interface with the database that will yield the
lat/long of the image, and then writing that back to the TIFF (now that
I know something about GeoTIFF, probably as a GeoTIFF now instead of
TIFF).  It looks like listgeo/geotifcp are the tools I would want to be
using for doing that.

So, I am just asking for some guidance from folks who have used these
tools and could comment on what might work well for what I am looking
for.  If you want more details, let me know.

Thanks,
Justin




_______________________________________________
Geotiff mailing list
[hidden email]
http://lists.maptools.org/mailman/listinfo/geotiff


_______________________________________________
Geotiff mailing list
[hidden email]
http://lists.maptools.org/mailman/listinfo/geotiff
Reply | Threaded
Open this post in threaded view
|

Re: Image processing project - getting started

David Burken
Justin,

If you have access to the actual sensor/camera physics/support data, one way to do the initial geo-positioning would be to create a sensor model and then orthorectify the image to the terrain.   You could do this with the ossim library. 

See: http://trac.osgeo.org/ossim/

The tricky part would be making the sensor model if one does not exist already.  If you can't do that Kyle's suggestion might be best, perhaps using gdalwarp.

Take care,
Dave


On 11/08/2010 02:26 PM, Kyle Shannon wrote:
Justin,
I would look at GDAL, gdal.org.  There are a few things that may be of use for you.  There are several command line utilities that may be of use and good a good api for C/C++, python and other languages.  The documentation is good too.  Read through the website and see if it works for you.

kss

# ============================
Kyle Shannon
Physical Science Technician
RMRS Fire Sciences Lab
Fire, Fuels & Smoke - RWU 4405
5775 Highway 10 W.
Missoula, MT 59808
(406)829-6954
[hidden email]
# ============================


On Mon, Nov 8, 2010 at 10:32 AM, Justin Close <[hidden email]> wrote:
Hello all.  I am new to the list and I am looking for some guidance and
pointers.

I am a computer science student working on a project involving
manipulation of a large number of images.  The images are aerial shots
of the cities at night (taken from space, actually).  We are taking the
RAW camera images, saving to TIFF, using the date/time stamp of the
image to generally locate what/where the image is of, then manually
morphing the image to scale/rotate/position it accurately to reality.
Reality in this case is Google Earth; the images will eventually be
served up as Google Earth layers.

The CS part of this is how to automate this process.  We wanted to embed
the lat/long information we got back into the image, and GeoTIFF seems
to be the perfect format for this.  Low and behold, it can also contain
morphing information too, it seems.  And there are all these nifty tools
out there that can do this manipulation, apparently.  But for me, who is
just starting out with it all, it seems a bit overwhelming; I don't know
where to start in order to find out what tools I should be using.
Libgeotiff seems like it has a bunch of things that we (it is a team
project) should be using; we don't want to reinvent the wheel (not
completely anyway - a bit of learning by doing can be good, though).

So far the images have just been processed within GE itself, using its
Image overlay features.  But that is tedious and awkward.  First step is
perhaps a better manual tool that can achieve the same results (an image
that is rotated and scaled to fit GE accurately).  A next step is to try
and get those manual operations automated so that a large number of
images can be done quickly, without intervention.

For the tagging (approximate lat/long) I was thinking of something to
read the TIFF tag data, interface with the database that will yield the
lat/long of the image, and then writing that back to the TIFF (now that
I know something about GeoTIFF, probably as a GeoTIFF now instead of
TIFF).  It looks like listgeo/geotifcp are the tools I would want to be
using for doing that.

So, I am just asking for some guidance from folks who have used these
tools and could comment on what might work well for what I am looking
for.  If you want more details, let me know.

Thanks,
Justin




_______________________________________________
Geotiff mailing list
[hidden email]
http://lists.maptools.org/mailman/listinfo/geotiff

_______________________________________________ Geotiff mailing list [hidden email] http://lists.maptools.org/mailman/listinfo/geotiff

_______________________________________________
Geotiff mailing list
[hidden email]
http://lists.maptools.org/mailman/listinfo/geotiff