[HOT] Suggestions for Helicopter-based Imagery

Stewart Long stewart at gonzoearth.com
Mon Feb 14 00:57:43 GMT 2011


Hi, very late to this thread and Nicholas may have already flown this last
Friday, but I have a few thoughts. The suggestions already given by others
look great.
Miko is correct about a prime lens/fixed focal length being most optimal.
The Canon point and shoot cameras that you do mention will work OK if that
is all you have access too.
I could help you do the full calculation for those two Canons, but as a
point of reference here is a typical Swath calculation from a Cannon G10


  *MODEL*

*ELEVATION in M*

*SWATH in M*

*SURFACE  Ha*

*RESOLUTION  CM*

*CANON G10*

100

124.590163934426

1.16420317119054

2.8

200

249.180327868853

4.65681268476216

5.6

300

373.770491803279

10.4778285407149

8.5

500

622.950819672131

29.1050792797635

14.1

1000

1245.90163934426

116.420317119054

28.2

1500

1868.85245901639

261.945713517872

42.3


 I am not sure what you spatial resolution requirements are, but the higher
you fly, the easier it will be to process the imagery. Holding the camera in
a consistent geodata position is critical. The bubble level should help
dampen vibrations. I simple hold onto the rubber clamp to reduce the
vibrations. Depending on the type of heli, it may be possible to use the
same simple rubber clamp configuration. This clamp is easy to set up, but it
really depends on the type of aircraft and coordination with the pilot.

Sufficient overlap amongst the images makes for a better map, but if you are
georeferencing and rectifying the images while focusing on the ground of the
scene, it is OK to include oblique perspectives on 3d objects like buildings
and trees. You just need to make sure that the actual ground is positioned
correctly. In an optimal situation there is >80% overlap, and a subset of
each image is made to crop out area that are off the vertical axis. If there
is less data, or you would like to include more information with the oblique
stuff on the periphery of the imagery that is fine too, keeping in mind that
you are correcting for the ground and not vertical objects.

Good Luck!
Stewart
-let me know if you have more questions, or if you want me to take a look at
the raw data for processing...



On Thu, Feb 10, 2011 at 1:04 AM, nicolas chavent
<nicolas.chavent at gmail.com>wrote:

> Hi there-
>
> Great to see the development of this thread, just a thought here: it might
> be intresting to involve persons from the OpenAerialMap<http://www.openaerialmap.org/Main_Page>(OAM) project where those topice have been and are being discussed to
> maximise synergies, this also calls for articulating the efforts of both
> group when tackling production/processing/use of imagery in crisis responses
> when the imagery is not served to osm directly the way yahoo, bing or
> spot/unosat did in the past.
>
> Excellent day to all
> Nicolas
>
>
> On Thu, Feb 10, 2011 at 8:00 AM, Milo van der Linden <milo at dogodigi.net>wrote:
>
>> One thing I noticed from Jeffreys pictures of the simple rig is the tape
>> to keep the lens in fixed position: there are lenses available with "fixed
>> focal length", these would be best to be used for aerial photography! And
>> the lenses are only a tiny bit more expensive then regular SLR lenses.
>>
>> hugin is indeed a good tool for this kind of work, but hugin results
>> depend largely on the "known variables", the more known variables you have,
>> the better it is.
>>
>> I was once the secretary of the dutch opengeo stichting (blog.opengeo.nl)
>> we had a RC helicopter project running. The fact that the project leader
>> failed to publish why aerial photography with rc helicopters was (not)
>> possible but simply said: they cannot be used, we will move to building a RC
>> plane was amongst one of the reasons for me to quit opengeo.nl
>>
>> If you take this thing serious and want to set up something general, not
>> just for one area; I would gladly introduce you to a couple of contacts of
>> mine at ITC Enschede http://www.itc.nl/ a university specialized in
>> GeoScience for developing countries.
>>
>>
>> General ideas:
>>
>> 1. Use a camera with fixed focal length lens (or prime lens)
>> 2. Create as much ground control points as possible. Don't forget to take
>> height into account.
>> 3. Calibrate the camera. Make sure you get the REAL focal length plus all
>> other parameters as correct as possible as opposed to what the manufacturer
>> gave as focal length
>> 4. Try to do as much as possible with hugin, their tool is excelent and
>> their community is eager to help out when you can clearly state requirements
>> regarding ortho photography
>>
>> [1] http://en.wikipedia.org/wiki/Prime_lens
>> [2] http://www.geo.u-szeged.hu/~joe/fotogrammetria/gcp_specs.htm
>> [3] http://hugin.sourceforge.net/tutorials/calibration/en.shtml
>> [4] http://hugin.sourceforge.net/
>>
>>
>>
>> 2011/2/10 Jeffrey Warren <jeff at unterbahn.com>
>>
>> What's interesting about that technique is that it uses Hugin to
>>> orthorectify imagery but he mentions that he is manually correcting ground
>>> control points (GCPs). This almost invalidates the use of hugin, but makes
>>> sense since what you really want to do is rectify against a reference map,
>>> so that the data doesn't drift.
>>>
>>> Manual stitching has for the most part proved to be much more accurate,
>>> even for large areas, and especially for relatively oblique imagery. (and in
>>> fact because in the osm.totor.ph example, he is actually manually moving
>>> GCPs, he might be better off using a program which allows a reference layer)
>>>
>>> Jeff
>>>
>>>
>>> On Thu, Feb 10, 2011 at 12:55 AM, maning sambale <
>>> emmanuel.sambale at gmail.com> wrote:
>>>
>>>> An osmer used this approach in the Philippines:
>>>> http://osm.totor.ph/
>>>>
>>>> On Thu, Feb 10, 2011 at 1:31 PM, george chamales <george at konpagroup.com>
>>>> wrote:
>>>> > Hello everyone,
>>>> > I'm looking for tips / recommendations / suggestions for the ideal way
>>>> to
>>>> > collect and process photos taken from a helicopter (full-size, not RC)
>>>> for
>>>> > use in OSM.
>>>> > At the very least we'll have a decent point-and-shoot, a gps, and a
>>>> > helicopter.  Still waiting to hear what our ceiling is and am trying
>>>> to hunt
>>>> > down a bubble level.
>>>> > Would be great to get some information back soon - there's a
>>>> possibility
>>>> > that we could be in the air as soon as Friday.
>>>> > george
>>>> > George Chamales
>>>> > Konpa Group
>>>> > Mobile: +1 718.288.7718
>>>> > Fax:  +1 857.488.4002
>>>> > Skype:  notgeorge
>>>> > http://konpagroup.com
>>>> >
>>>> > _______________________________________________
>>>> > HOT mailing list
>>>> > HOT at openstreetmap.org
>>>> > http://lists.openstreetmap.org/listinfo/hot
>>>> >
>>>> >
>>>>
>>>>
>>>>
>>>> --
>>>> cheers,
>>>> maning
>>>> ------------------------------------------------------
>>>> "Freedom is still the most radical idea of all" -N.Branden
>>>> wiki: http://esambale.wikispaces.com/
>>>> blog: http://epsg4253.wordpress.com/
>>>> ------------------------------------------------------
>>>>
>>>> _______________________________________________
>>>> HOT mailing list
>>>> HOT at openstreetmap.org
>>>> http://lists.openstreetmap.org/listinfo/hot
>>>>
>>>
>>>
>>> _______________________________________________
>>> HOT mailing list
>>> HOT at openstreetmap.org
>>> http://lists.openstreetmap.org/listinfo/hot
>>>
>>>
>>
>> _______________________________________________
>> HOT mailing list
>> HOT at openstreetmap.org
>> http://lists.openstreetmap.org/listinfo/hot
>>
>>
>
>
> --
> Nicolas Chavent
> Humanitarian OpenStreetMap Team
> http://wiki.openstreetmap.org/wiki/WikiProject_Haiti
> Mobile (Haiti): +509 389 583 05
> Mobile (France): +33 6 89 45 54 58
> Landline (FRA): +33 2 97 26 23 08
> Email: nicolas.chavent at gmail.com
> Skype: c_nicolas
> Twitter: nicolas_chavent
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/hot/attachments/20110213/9aa4934e/attachment-0001.html>


More information about the HOT mailing list