[HOT] Suggestions for Helicopter-based Imagery

Milo van der Linden milo at dogodigi.net
Thu Feb 10 14:32:26 GMT 2011

> Questions:
> * Are there any general rules for relating megapixels / aspect ratio
> to altitude for getting good pictures?

In practice; 5Megapixel is sufficient from a UAV at max. 100m height, but
would do up to 200 meters.

> * If we have GCPs, do we need to have gps coords for each picture?
It would be good if you indeed have the GPS position of the heli when the
picture was taken. Again, it would be good to calibrate the relative
position of the camera to the helicopters GPS and make sure that camera
clock matches the GPS clock. If you don't have GPS positions for the camera,
make sure you have enough overlap between seperate images. 60% overlap is
used to generate stereopairs but will also help you to eliminate
camera-position errors.

* What's the minimum number of GCPs we need?

> * Are there any best practices for collecting GCPs?

I could ask my scientific contacts at ICT Enschede.

> * What programs should we be looking at to do orthorectification and
> is there another mailing list that's filled with orthorectification
> ninjas who would love to help out?

Use hugin, it will give you the best orthorectification for "free". If you
want to use professional software; 10.000 euro per license is not enough....
You have to decide what quality you need at what price. Centimeter accuracy
is nonsense, with the way you propose gathering photo's you should be happy
with a meter accuracy which will perfectly fit the purpose in my opinion.
Another GIS tool tailored for image processing is ILWIS
http://www.itc.nl/news_events/archive/research/0011.asp it has become open
source, windows only. Visit the page for more info.

If you find ways to generate a good first set of images, I will be glad to
help (remotely) to find out what steps in the process should be fine tuned,
enhanced and so on. I would go for a "trial and error" process here, make a
first flight, see what we can do with it, fine tune, make a second flight,
evaluate and so on... this is practical, will not cost to much money and
will not cause delay. Worst case you will have photographs from the air that
have to be interpreted by human eyes and turned in to mapdata, Best case
will give you perfectly alligned hig quality imagery that can be used in
mapping/gis software. You have nothing to lose!

> Thanks for the info,
> george
> On 2/10/11, Milo van der Linden <milo at dogodigi.net> wrote:
> > If I remember correctly; there is currently no real activity in the
> > openaerialmap project. There are however some things going on at
> > https://github.com/oam/oam although the last commits seems to be from
> > november 2010.
> >
> > I think it would be good to contact the people at OAM and ask them what
> > their plans are first.
> >
> > 2011/2/10 Matthias Meißer <digi_c at arcor.de>
> >
> >> I'm not an expert but you might be lucky with 2 cams and doing
> >> http://wiki.openstreetmap.org/wiki/Photogrammetry
> >> if your area has different highlevels and your imagery could be
> >> disordered.
> >>
> >> AFAIK the most would preffer quadcopter drones cause their flight is
> much
> >> more stable (that are unavailable for you I guess) but here is a nice
> >> video
> >> showing you the principles:
> >> http://www.mikrokopter.de/ucwiki/VideoAbspielen?id=206
> >>
> >> regards
> >> Matthias
> >>
> >>
> >> _______________________________________________
> >> HOT mailing list
> >> HOT at openstreetmap.org
> >> http://lists.openstreetmap.org/listinfo/hot
> >>
> >
> --
> George Chamales
> Rogue Genius LLC
> Phone:  718.288.7718
> Skype:  notgeorge
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/hot/attachments/20110210/e6c034c1/attachment.html>

More information about the HOT mailing list