[HOT] Spot-6 images of Guinée / Northern Liberia available, job 489

Jean-Guilhem Cailton jgc at arkemie.com
Sat Apr 5 20:17:55 UTC 2014


Hi Michael,

Le 05/04/2014 21:25, Michael Judex a écrit :
> Thanks for the answer! I've another related question/issue: There is a
> systematic geometric shift between SPOT-6 images of task 489 (Guinée /
> Northern Liberia border area) and Pleiades-1A images of task 471 (only
> for the town of Macenta, which is included geographically in 489). I
> think that an offset of 10.44; 14.24 (in JOSM) for the SPOT-6 images
> is good to have a proper fit to the Pleiades images.
>
> So my questions are:
> - is the geometric accuracy of the Pleiades better than SPOT-6? (I
> assume yes)

Yes, Pleiades has 6.5 m CE90 location accuracy, and SPOT-6 10 m CE90
location accuracy.

> - do we then have a systematic geometeric distortion in task 489 for
> the edits that relied only on the SPOT-6 imagery? If yes can we
> correct this? (maybe the the only way to really fix offsets between
> different images are GPS ground control points...)

The shift in Macenta doesn't mean that the same offset there would be
suitable for the whole SPOT-6 imagery. It is likely suitable only for
the area near Macenta.

Geometric distortions are for example likely to depend on relief, in a
way more complex than a simple uniform offset could fix.

>
> I think it could be helpful to have an instant overview of ALL
> available imagery for a specific area. If a mapper opens a task only
> the dedicated imagery is listed in the task instructions. Is there a
> simple way to achieve this?

You can simply overlay the Pleiades imagery for one of the Pleiades jobs
(469 or 471 for Guéckédou or Macenta - both contain the same Pleiades
layer, with Kissidougou too ; and the Pleiades imagery layer for job
http://tasks.hotosm.org/job/487 for Foya contains the 4 Pleiades images)
to the SPOT-6 imagery of job http://tasks.hotosm.org/job/489. And switch
them on and off to compare (which is what you probably did, explaining
here for the others).


A better answer to your question : on VISOV's Umap used to track news
about the Ebola outbreak, boundaries for high resolution imagery layers
can be switched on and off (in the layer menu, icon is near top left)
(grouped by when they where made available) :

http://umap.openstreetmap.fr/fr/map/esov-suivi-epidemie-ebola_6356#10/8.5640/-9.7806


>
> Thanks for your support!
>
> Michael


Thanks for your questions, and for your mapping!

Jean-Guilhem


>
>
>
> Am 05.04.2014 11:19, schrieb Pierre Béland:
>> Great. Thanks both.
>>  
>> Pierre
>>
>> ------------------------------------------------------------------------
>> *De :* Jean-Guilhem Cailton <jgc at arkemie.com>
>> *À :* Michael Judex <m.judex at gmx.de>; hot at openstreetmap.org
>> *Envoyé le :* Vendredi 4 avril 2014 22h10
>> *Objet :* Re: [HOT] Spot-6 images of Guinée / Northern Liberia
>> available, job 489
>>
>> Hi Michael,
>>
>> The existing roads were probably based on Landsat, the geometric
>> accuracy of which is lower than Spot-6. So if you feel that adjustment
>> is needed, the pre-existing road should probably be adjusted. This would
>> also make it possible to have a consistent geometry for tracing streets
>> and buildings within villages and towns.
>>
>> Best wishes,
>>
>> Jean-Guilhem
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/hot/attachments/20140405/b52a3dc0/attachment.html>


More information about the HOT mailing list