Determining DEM Resolution

[Nolan Fewell]

Hi all, I made several CTX DEMs for a project, and I have to report on their pixel size and spatial resolution.  I was able to simply use the measuring tool in ArcMap to measure the size of a pixel, but I was told that might not be completely accurate. Is there another way to determine that information?

Thanks,

Nolan

[Scott McMichael]

gdalinfo should show you the pixel size of a DEM file:

$ gdalinfo V03_highres.tif
Driver: GTiff/GeoTIFF
Files: V03_highres.tif
Size is 2808, 2816
Coordinate System is:
PROJCS["WGS 84 / Antarctic Polar Stereographic",
GEOGCS["WGS 84",
DATUM["WGS_1984",
SPHEROID["WGS 84",6378137,298.2572235604902,
AUTHORITY["EPSG","7030"]],
AUTHORITY["EPSG","6326"]],
PRIMEM["Greenwich",0],
UNIT["degree",0.0174532925199433],
AUTHORITY["EPSG","4326"]],
PROJECTION["Polar_Stereographic"],
PARAMETER["latitude_of_origin",-71],
PARAMETER["central_meridian",0],
PARAMETER["scale_factor",1],
PARAMETER["false_easting",0],
PARAMETER["false_northing",0],
UNIT["metre",1,
<---------------------------
AUTHORITY["EPSG","9001"]],
AUTHORITY["EPSG","3031"]]
Origin = (-2399659.404000000096858,1099578.827000000048429)
Pixel Size = (8.000000000000000,-8.000000000000000)
<-------------------------------------
Metadata:
AREA_OR_POINT=Point
Image Structure Metadata:
INTERLEAVE=BAND
Corner Coordinates:
Upper Left (-2399659.404, 1099578.827) ( 65d22'54.00"W, 66d 2'49.48"S)
Lower Left (-2399659.404, 1077050.827) ( 65d49'40.06"W, 66d 7'45.04"S)
Upper Right (-2377195.404, 1099578.827) ( 65d10'37.05"W, 66d13'37.76"S)
Lower Right (-2377195.404, 1077050.827) ( 65d37'33.05"W, 66d18'35.83"S)
Center (-2388427.404, 1088314.827) ( 65d30'11.03"W, 66d10'42.73"S)
Band 1 Block=2808x1 Type=Float32, ColorInterp=Gray
NoData Value=-32767
Unit Type: metre


Scott

> --
> You received this message because you are subscribed to the Google Groups "Ames Stereo Pipeline Support" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to ames-stereo-pipeline...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/ames-stereo-pipeline-support/636f0599-0a97-44be-971d-2c5ef5395aff%40googlegroups.com.

[Nolan Fewell]

Thank you!!

> To unsubscribe from this group and stop receiving emails from it, send an email to ames-stereo-pipeline-support+unsub...@googlegroups.com.

[Etienne Berthier]

Hello Nolan,

gdalinfo provide the DEM pixel size or the ground sampling distance. Note that this is different from the horizontal resolution of the dataset.

Starting from an initial dataset of a given resolution you can always oversample or downsample it using various interpolators. Doing so, you change the pixel size (=ground sampling distance). But you do not change the resolution itself i.e. the level of details that can be observed using the dataset.

See in particular https://topex.ucsd.edu/sandwell/publications/97.pdf

--

You received this message because you are subscribed to the Google Groups "Ames Stereo Pipeline Support" group.

To unsubscribe from this group and stop receiving emails from it, send an email to ames-stereo-pipeline...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ames-stereo-pipeline-support/636f0599-0a97-44be-971d-2c5ef5395aff%40googlegroups.com.

--

Etienne Berthier           

CNRS / OMP / LEGOS                            http://etienne.berthier.free.fr/       
14 av. Ed. Belin 31400 Toulouse        +33 5 61 33 47 15

[Alexandrov, Oleg (ARC-TI)[SGT, INC]]

The discussion of what is the effective resolution of a DEM, meaning the smallest feature that can be resolved, is an interesting one to have. In my experience, with our regular block matching stereo (--stereo-algorithm 0, with any subpixel method), the effective grid size of the DEM should be about 4X the ground sample distance of the images. MGM/SGM does provide more detail, then perhaps it becomes 2X the image GSD. With shape-from-shading on the Moon (when one does a careful job) it could maybe even go to closer to 1X.

Note that point2dem does not interpolate when creating a DEM, as that would mean that out of all points in the cloud, only the nearest to the output grid of the DEM contribute (whichever interpolation method is used). Instead, for each output grid point, it averages all the cloud points within about 1/2 to 1 grid size distance from it. The precise distance as well as the weight used in averaging (the sigma of the Gaussian) are customizable. Hence, each point cloud grid point contributes to the final DEM even if that one has a coarser grid than the cloud itself. 

---

From: ames-stereo-pi...@googlegroups.com <ames-stereo-pi...@googlegroups.com> on behalf of Etienne Berthier <etienne....@legos.obs-mip.fr>
Sent: Thursday, August 22, 2019 11:47 PM
To: Fewell, Nolan P (JPL-3223)[Affiliate] <nolan.p...@jpl.nasa.gov>
Cc: Ames Stereo Pipeline Support <ames-stereo-pi...@googlegroups.com>
Subject: [EXTERNAL] Re: Determining DEM Resolution

 

To view this discussion on the web visit https://groups.google.com/d/msgid/ames-stereo-pipeline-support/ddc33b6d1acbce6eabd68a709dffe5af%40legos.obs-mip.fr.