Jump to content

 
Photo

ASTER DEM

- - - - -

  • Please log in to reply
10 replies to this topic

#1
david17tym

david17tym

    Key Contributor

  • Validated Member
  • PipPipPip
  • 99 posts
  • Gender:Male
  • Location:North Wales
  • United Kingdom

Hello all,

I downloaded some tiles covering North Wales, gridded and merged them in ArcGIS and hillshaded the result with 3D Analyst. All fine so far - but taking a closer look (see attachment) I've got some curious worm like features appearing.

Am I doing something wrong with the data, only as a native of these parts I don't recall seeing anything like this on the ground.

Snowdon is at bottom left.

Thanks,

Dave

Attached Files



#2
rudy

rudy

    Ultimate Contributor

  • Validated Member
  • PipPipPipPipPipPip
  • 739 posts
  • Gender:Male
  • Location:Canada
  • Canada

See the previous discussion of ASTER here. I suspect it is a data problem rather than a process or software problem. I haveb't had a chance yet to look at some of the data closely yet.

#3
Sergio Huykman

Sergio Huykman

    Contributor

  • Validated Member
  • PipPip
  • 11 posts
  • Gender:Male
  • Location:Argentina
  • Argentina

Hello all,

I downloaded some tiles covering North Wales, gridded and merged them in ArcGIS and hillshaded the result with 3D Analyst. All fine so far - but taking a closer look (see attachment) I've got some curious worm like features appearing.

Am I doing something wrong with the data, only as a native of these parts I don't recall seeing anything like this on the ground.

Snowdon is at bottom left.

Thanks,

Dave


Z factor too high, try with 0.000011559.

#4
david17tym

david17tym

    Key Contributor

  • Validated Member
  • PipPipPip
  • 99 posts
  • Gender:Male
  • Location:North Wales
  • United Kingdom

Hi Sergio,
I exaggerated the z-factor on purpose to illustrate my point.
The linear features are still visible when using the correct z-factor for the latitude.
I have contacted NASA/LP DAAC direct, hopefully they can shed some light on it.
Thanks,
Dave

#5
david17tym

david17tym

    Key Contributor

  • Validated Member
  • PipPipPip
  • 99 posts
  • Gender:Male
  • Location:North Wales
  • United Kingdom

Morning all,

According to NASA/LP DAAC the "mole run anomalies" are a recognised issue with the data
along with residual cloud and pit artifacts.

They also state ASTER GDEM Version 1 should be viewed as "experimental" or "research grade".

Either that or we have some HUGE wildlife roaming about........

Dave

#6
Hans van der Maarel

Hans van der Maarel

    CartoTalk Editor-in-Chief

  • Admin
  • PipPipPipPipPipPipPip
  • 3,868 posts
  • Gender:Male
  • Location:The Netherlands
  • Interests:Cartography, GIS, history, popular science, music.
  • Netherlands

Either that or we have some HUGE wildlife roaming about........


Graboids!
Hans van der Maarel - Cartotalk Editor
Red Geographics
Email: hans@redgeographics.com / Twitter: @redgeographics

#7
gregsd

gregsd

    Master Contributor

  • Validated Member
  • PipPipPipPip
  • 169 posts
  • Gender:Male
  • Location:Surrey, England.
  • Interests:Cycling, mostly off-road (MTB) riding.
  • United Kingdom

...Either that or we have some HUGE wildlife roaming about........

Dave


Might be clusters of sheep ;)


Greg Driver

GIS Analyst
MapInfo User...!

#8
Nick H

Nick H

    Legendary Contributor

  • Validated Member
  • PipPipPipPipPip
  • 307 posts
  • Gender:Male
  • Location:Caversham, Reading, England.
  • United Kingdom

Hi Dave, I've tried converting a couple of ASTER DEMs for my local area (the Thames Valley) to contour shape files using gdal_contour and the results are as disappointing as yours, if this is any consolation.

Regards, N.
Caversham, Reading, England.

#9
rudy

rudy

    Ultimate Contributor

  • Validated Member
  • PipPipPipPipPipPip
  • 739 posts
  • Gender:Male
  • Location:Canada
  • Canada

Either that or we have some HUGE wildlife roaming about........


Graboids!

Sandworms!

#10
liuba

liuba

    Newbie

  • New Member
  • Pip
  • 2 posts
  • Ukraine

Hello all,

I downloaded some tiles covering North Wales, gridded and merged them in ArcGIS and hillshaded the result with 3D Analyst. All fine so far - but taking a closer look (see attachment) I've got some curious worm like features appearing.

Am I doing something wrong with the data, only as a native of these parts I don't recall seeing anything like this on the ground.

Snowdon is at bottom left.

Thanks,

Dave


Unfortunately GDEM (ASTER DEM) is not as good as I thought in a beginning. It is a result of automatic optical stereo processing. SRTM at least has much less noisy data. Anyway I can’t use neither SRTM, no GDEM , cause normally I need the ground surface data (DTM) without influence of forest and build-up areas. But, definitely for demonstration, prior calculations and studying it’s a good data, because it’s free!

#11
ninja

ninja

    Newbie

  • Validated Member
  • Pip
  • 9 posts
  • United States

I've been using ASTER GDEM for an ortho image acquisition project in the Philippines recently. I immediately noticed that something was amiss. The imagery that looked ok with 90m SRTM DEM was now tearing apart at the seams. Duplication across images was the norm it seemed especially in area of high relief. Reading the 'validation report' confirmed this as a known issue. In the same report I read that the RMSE for the whole set is +/-10m in the vertical as compared to the NED set. Less than impressed I read on and the Japanese Ministry of Economy, Trade and Industry reported and RMSE of +/-5m! What?! How can that be? Turns out the ASTER GDEM is horribly geolocated too! To arrive at the better RMSE the Japanese registered the DEM before conducting their validation. Sooooo... "If it works for them... " I thought. I force registered one cell (1deg x 1deg) of the ASTER GDEM to the same origin as the same 1deg tile of SRTM 90m and reprocessed the image data.

The results were markedly better. I can not say that it's properly located but it's closer by an order of magnitude, indeed, with the worst of the problems now appearing in higher elevation water courses. You'll remember this area just recently experienced multiple typhoons. I have to wonder if those areas at least in some regard no longer agree with the DEM. Further, the particular area (west Luzon island) is heavily terraced and if DEM is to tree canopy as it says in the Validation summary, this could also be an influencing factor.

If anyone is in that spot I would recommend trying the technique I mention of forcing the registration to be the same as the SRTM. See if it helps.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

-->