[gdal-dev] re-gridding to a coarser grid

[Doug_Newcomb at fws.gov]

For what it's worth,

What do you consider large and what kind of method for downscaling you 
wish to use?

  I do this in GRASS using the r.neighbors command, 
http://grass.fbk.eu/grass64/manuals/html64_user/r.neighbors.html.   I've 
done an averaging method to convert a 20 ft elevation grid  ( statewide 
coverage)  of approximately 7 billion cells to an averaged 60 ft grid ( 
777 million cells)  by performing a 3x3 neighbor hood analysis  on the 20 
ft grid, setting the GRASS workspace resolution to 60 ft and multiplying 
the 20ft grid created by the  3x3 neighborhood analysis by 1 to create a 
new grid (
http://grass.fbk.eu/grass64/manuals/html64_user/r.mapcalculator.html)   . 
GRASS will take the value of the center averaged 20 ft cell in the 60ft 
block as the value of the cell in the new 60 ft raster. 

options for r.neigbors calculations are : 
average,median,mode,minimum,maximum,range,stddev,sum,variance,diversity,interspersion 



Doug


Doug Newcomb 
USFWS
Raleigh, NC
919-856-4520 ext. 14 doug_newcomb at fws.gov
---------------------------------------------------------------------------------------------------------
The opinions I express are my own and are not representative of the 
official policy of the U.S.Fish and Wildlife Service or Dept. of the 
Interior.   Life is too short for undocumented, proprietary data formats.



Tim Keitt <tkeitt at gmail.com> 
Sent by: gdal-dev-bounces at lists.osgeo.org
12/15/2011 01:17 PM

To
"Andreas H." <lists at hilboll.de>
cc
gdal-dev at lists.osgeo.org
Subject
Re: [gdal-dev] re-gridding to a coarser grid






If you need complete control, and your data are not too massive in
number, I have had good results pushing pixel coordinates as points
into postgis and then using spatial queries to aggregate in various
ways (eg averages a hexagonal grid). It seems round-about, but it
works. I've actually done it on some pretty massive grids.

THK

On Wed, Dec 14, 2011 at 7:26 AM, Andreas H. <lists at hilboll.de> wrote:
> Travis,
>
> thanks for your answer!
>
> Regarding the resampling methods: Do they all just interpolate the data? 
I
> mean, when downsampling, usually I would use mean() or something similar
> to fill the new (coarser) grid cells. Doas gdalwarp actually do this and
> I'm not able to understand the documentation, or is it different?
>
> Thanks again!
> Andreas.
>
>
>> Andreas,
>>
>> gdalwarp can be used to resample images using the -tr flag or -ts flag.
>>
>> For example resample 1m image to 10m using cubic resampling and
>> "target resolution'
>>
>> gdalwarp -r cubic -tr 10 10 input_1m.tif output_10m.tif
>>
>> You may have an issue determining the output resolution as I believe
>> the resolution will be in decimal degrees rather than arc-seconds.
>>
>> Regards
>>
>> On 13 December 2011 14:52, Andreas H. <lists at hilboll.de> wrote:
>>> Hi,
>>>
>>> let's say I have a GeoTIFF file with a global grid in a 30 arc-second
>>> resolution. Which would be the appropriate GDAL command to spatially
>>> down-sample this file to say 0.125°?
>>>
>>> Thanks for your insight,
>>> Andreas.
>>> _______________________________________________
>>> gdal-dev mailing list
>>> gdal-dev at lists.osgeo.org
>>> http://lists.osgeo.org/mailman/listinfo/gdal-dev
>>
>
>
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/gdal-dev



-- 
Timothy H. Keitt
http://www.keittlab.org/
_______________________________________________
gdal-dev mailing list
gdal-dev at lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/gdal-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/gdal-dev/attachments/20111215/364df1f6/attachment.html