orthorectification,aerial photographs

[William L. Baker]

Radoslav,
I've used i.ortho.photo many times.  You must
have a single scan of enough of the photo to 
include the fiducial marks, since the program
sets up a mathematical transformation from the
location of the fiducials, the control points on
the ground, and the camera characteristics.
After you generate the orthophoto, you can
trim out a smaller area if you like.  Just use
d.zoom or d.rast.zoom to zoom in to the area,
then r.resample to create a new raster map
containing only the zoomed area.

You can focus your efforts on the small area
containing the raster elevation model and only
generate control points for that area, but you
still must have the whole photo to work with,
including the fiducials.  If you focus on a small
area when you put in the control points, the
resulting orthophoto will be poor outside the
area of the control points. Good luck.

Bill Baker
Univ. of Wyoming

>----------
>From: 	Radoslav Bonk[SMTP:bonk at nic.fns.uniba.sk]
>Sent: 	Sunday, September 14, 1997 4:14 PM
>To: 	grassu at cecer.army.mil
>Subject: 	Re: orthorectification,aerial photographs
>
>Thank you for your advice it was very clearly.
>I have a small successes in my project
>- have raster elevation model of desired area, even the aerial
>photographs /4/ cover larger area, does it matter?
>- got one aerial photograph with r.in.tiff into GRASS, just for training,
>because about 0.5 cm of the photo is missing /part with fiducial marks/
>- have whole camera characteristics 
>- have done such as i.group, i.target ...
>
>my guestions: Is it necessary to scan whole aerial photo, isn't it?
>I'm trying to do more 
>Rado Bonk              
>    
>
>*************************************************
> Radoslav Bonk
> Comenius University,Faculty of Natural Sciences
> Department of Physical Geography and Geoecology
> Mlynska Dolina 842 14,Bratislava SLOVAKIA
> e-mail: bonk at nic.fns.uniba.sk
> PRIVAT: Svinia 221, 082 32; tel:+42/91/958 187
>*************************************************
>
>
>On Sun, 14 Sep 1997, Schmidt wrote:
>
>> I've had good success using r.in.gif for bringing scanned aerial photos
>>into
>> GRASS and using i.ortho.photo to perform the ortho corrections. To do the
>> corrections for relief distortion, you will need a GRASS raster elevation
>> map for the area. You will also need to input information on the aerial
>> camera characteristics (focal length and fiducial measurements). You will
>> then need to record the locations of the fuducial marks on the scanned
>> photo, register the photo to some other rectified image or map, and then
>> rectify it.  
>> 
>> If you have further questions, I will try to help.
>> 
>> Jim Schmidt
>> avenues at sonnet.com
>> 
>> 
>> 
>> 
>> 
>> 
>> At 06:49 PM 9/8/97 +0200, you wrote:
>> >>
>> >>Hi my name is Radoslav Bonk
>> >>I'm working on the project where I need to put an aerial photographs
>> >>strictly on a topographical map for solving the LANDUSE. It's a hilly
>> >>region and it gets a quite big problems even I worked in EASI/PACE.
>> >>I have no other softwares except the GRASS's tools for orthorectifying
>> >>aerial photos. I need the photos orthorectify and then try to
>> >>georeference them as sharp as possible on topographic maps.
>> >>I have no experiences with image processing in GRASS, even i did mostly
>> >>morphometric modeling and r.mapcalc modeling of soil erosion and other
>> >>geomorphological hazards.
>> >>
>> >>my GUESTION is: 1. Are the GRASS tools for orthorectifying an aerial
>> >photos suficient to get a good results?
>> >                 2. What are the steps of such an image processing?
>> >                 3. I have compiled r.in.tiff I think I will need it for
>> >image processing. Will I?
>> >                 4. Where I can find a similar project or www paper with
>> >>    solving the problems of Orthorectifying the aerial photos?
>> >
>> >********************************************************************
>> >I would like to ask somebody for consulting this problems on the net.
>> >********************************************************************
>> 
>> >>
>> >>Thank you very much
>> >>                                                 Rado Bonk
>> >> 
>> >>*************************************************
>> >> Radoslav Bonk
>> >> Comenius University,Faculty of Natural Sciences
>> >> Department of Physical Geography and Geoecology
>> >> Mlynska Dolina 842 14,Bratislava SLOVAKIA
>> >> e-mail: bonk at nic.fns.uniba.sk
>> >> PRIVAT: Svinia 221, 082 32; tel: 091/958 187
>> >>*************************************************
>> >>
>> >
>> >
>> avenues at sonnet.com
>> 
>
>