From schuyler at nocat.net  Wed Feb  3 08:33:59 2010
From: schuyler at nocat.net (Schuyler Erle)
Date: Wed, 3 Feb 2010 10:33:59 -0500
Subject: [OAM-talk] Fwd: [Geowanking] Interesting article on Open Data
	(crossposted)
Message-ID: <20100203153359.GA7306@vishnu.tridity.org>

Bears some thinking about, esp. w/r/t OAM. In a nutshell, posting your
code on Sourceforge is a very different thing from building a F/OSS
dev community; Nat asserts the same is true for data. Great read.

SDE

----- Forwarded message from Ian Turton <ijturton at gmail.com> -----

Date: Tue, 2 Feb 2010 16:09:47 -0500
From: Ian Turton <ijturton at gmail.com>
To: Geowanking <Geowanking at geowanking.org>
Subject: [Geowanking] Interesting article on Open Data

Rethinking Open Data:Lessons learned from the Open Data front lines by
Nat Torkington
http://radar.oreilly.com/2010/02/rethinking-open-data.html

apparently you can build an open data site and users will not come!

Ian

-- 
Ian Turton

Sent from State College, PA, United States

_______________________________________________
Geowanking mailing list
Geowanking at geowanking.org
http://geowanking.org/mailman/listinfo/geowanking_geowanking.org

----- End forwarded message -----


From warren at mit.edu  Wed Feb 17 09:36:03 2010
From: warren at mit.edu (Jeffrey Warren)
Date: Wed, 17 Feb 2010 11:36:03 -0500
Subject: [OAM-talk] open source orthorectification for grassrootsmapping.org
Message-ID: <f3515e331002170836o1a1d44e6y317283a922121be8@mail.gmail.com>

Hi all - I'm looking for an open source way to do automated
orthorectification as a web service to support
http://grassrootsmapping.org balloon/kite
mapping imagery.

Something that would generate GCPs and stitch imagery, and (even better)
perhaps generate tiles.

I'm a bit new to rectification, any suggestions?

Thanks,
Jeff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://host134.hostmonster.com/pipermail/talk_openaerialmap.org/attachments/20100217/6e45fec5/attachment.html>

From MPfister at enplan.com  Wed Feb 17 12:40:43 2010
From: MPfister at enplan.com (Marc Pfister)
Date: Wed, 17 Feb 2010 11:40:43 -0800
Subject: [OAM-talk] open source orthorectification for
 grassrootsmapping.org
In-Reply-To: <f3515e331002170836o1a1d44e6y317283a922121be8@mail.gmail.com>
References: <f3515e331002170836o1a1d44e6y317283a922121be8@mail.gmail.com>
Message-ID: <BBE7E7E30E9D9B42B16A75F3C8823D1D32CF1B35@SBSRAQ.enplan.local>

There's some or all of the functionality in GRASS, OSSIM, E-FOTO, and the NASA Vision Workbench. Making it all work as a webservice sounds tough to me.

I think you'll also have some challenges it that you probably only have low resolution and quite possibly low accuracy orthoimagery to control against, coarse (if any) DEMs, and a camera with a short focal length, no fiducials or calibration report. I wonder if simple warping will get you just about as far in terms of accuracy.

At Random Hacks of Kindness I started on some code that used SIFT to try to match UAV photos to control imagery and generate a GCP list that could be fed into gdalwarp. The UAV imagery did have some rough GPS positioning that limited the search space for the SIFT algorithm. It's still (argh) unfinished, but you can see the start here:

http://github.com/drwelby/UAV-Ortho

-Marc Pfister



From: talk-bounces at openaerialmap.org [mailto:talk-bounces at openaerialmap.org] On Behalf Of Jeffrey Warren
Sent: Wednesday, February 17, 2010 8:36 AM
To: geodata at lists.osgeo.org; talk at openaerialmap.org
Subject: [OAM-talk] open source orthorectification for grassrootsmapping.org

Hi all - I'm looking for an open source way to do automated orthorectification as a web service to support http://grassrootsmapping.org balloon/kite mapping imagery.

Something that would generate GCPs and stitch imagery, and (even better) perhaps generate tiles.

I'm a bit new to rectification, any suggestions?

Thanks,
Jeff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://host134.hostmonster.com/pipermail/talk_openaerialmap.org/attachments/20100217/e24d98ed/attachment.html>

From warren at mit.edu  Wed Feb 17 13:01:19 2010
From: warren at mit.edu (Jeffrey Warren)
Date: Wed, 17 Feb 2010 15:01:19 -0500
Subject: [OAM-talk] open source orthorectification for
	grassrootsmapping.org
In-Reply-To: <BBE7E7E30E9D9B42B16A75F3C8823D1D32CF1B35@SBSRAQ.enplan.local>
References: <f3515e331002170836o1a1d44e6y317283a922121be8@mail.gmail.com>
	<BBE7E7E30E9D9B42B16A75F3C8823D1D32CF1B35@SBSRAQ.enplan.local>
Message-ID: <f3515e331002171201r4f82fc4di993ad6f8d5f82da1@mail.gmail.com>

I started with Map Warper (warper.geothings.net) but i found that manually
warping in photoshop (just using the distort tool) yielded much better
results, and was actually faster. This is mainly because of the kinds of
images (and scales) you get from balloon mapping. See:

http://www.flickr.com/photos/jeffreywarren/sets/72157623196104469/

Thanks for the link - i'll take a look. I'd been using SIFT but haven't
fiddled with it enough to know how to rectify against control imagery. It
does a pretty good job for stitching, of course - though it doesn't
accommodate images at different scales that well - something which happens a
lot with balloon imagery.

<http://www.flickr.com/photos/jeffreywarren/sets/72157623196104469/>Best,
Jeff

On Wed, Feb 17, 2010 at 2:40 PM, Marc Pfister <MPfister at enplan.com> wrote:

>  There?s some or all of the functionality in GRASS, OSSIM, E-FOTO, and the
> NASA Vision Workbench. Making it all work as a webservice sounds tough to
> me.
>
>
>
> I think you?ll also have some challenges it that you probably only have low
> resolution and quite possibly low accuracy orthoimagery to control against,
> coarse (if any) DEMs, and a camera with a short focal length, no fiducials
> or calibration report. I wonder if simple warping will get you just about as
> far in terms of accuracy.
>
>
>
> At Random Hacks of Kindness I started on some code that used SIFT to try to
> match UAV photos to control imagery and generate a GCP list that could be
> fed into gdalwarp. The UAV imagery did have some rough GPS positioning that
> limited the search space for the SIFT algorithm. It?s still (argh)
> unfinished, but you can see the start here:
>
>
>
> http://github.com/drwelby/UAV-Ortho
>
>
>
> -Marc Pfister
>
>
>
>
>
>
>
> *From:* talk-bounces at openaerialmap.org [mailto:
> talk-bounces at openaerialmap.org] *On Behalf Of *Jeffrey Warren
> *Sent:* Wednesday, February 17, 2010 8:36 AM
> *To:* geodata at lists.osgeo.org; talk at openaerialmap.org
> *Subject:* [OAM-talk] open source orthorectification for
> grassrootsmapping.org
>
>
>
> Hi all - I'm looking for an open source way to do automated
> orthorectification as a web service to support
> http://grassrootsmapping.org balloon/kite mapping imagery.
>
>
>
> Something that would generate GCPs and stitch imagery, and (even better)
> perhaps generate tiles.
>
>
>
> I'm a bit new to rectification, any suggestions?
>
>
>
> Thanks,
>
> Jeff
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://host134.hostmonster.com/pipermail/talk_openaerialmap.org/attachments/20100217/a4ecf5a7/attachment.html>

From jrp.crs at gmail.com  Thu Feb 18 08:50:44 2010
From: jrp.crs at gmail.com (John Robert Peterson)
Date: Thu, 18 Feb 2010 15:50:44 +0000
Subject: [OAM-talk] Simple sift and stitch process?
Message-ID: <a336d45e1002180750t10adf71cgeab9cfabd66444c1@mail.gmail.com>

A number of times now the question has come up, and a some people have
partially answered it (including myself), but if the full answer has
come up, I have missed it:

If someone wants to take a bunch of images that are mostly rectified
and well behaved, sift them, and stich them together, what's the best
technique?

Hugin can do this awesomely, but as far as I know, insists on
surface-of-sphere comprehension (which is ironic given how spherical
the earth is, but that's not really the point). Is there a way to do
this flat, without jumping through too many hoops?

If noone has the complete answer, what's the closest we do have?

Thanks,
JR