[gdal-dev] Cassini_Soldner warping using WGS84 GCPs is distorted

Frank Warmerdam warmerdam at pobox.com
Fri Sep 18 10:39:09 EDT 2009


Martin Geng wrote:
> I have a city map in Cassini_Soldner projection with a separate file
> containing the Soldner coordinates of the 4 corners.
> I want to reproject it to WGS84 in order to render it e.g. with SharpMap and
> display some GPS positions.
> I have tried it with some different parameter settings of gdalwarp but in
 > every case the resulting image is distorted (the size of the generated tif is
> decreased to 15141, 6943).
> Even if with the parameter -ts 15118 11339 the original size is adjusted
> SharpMap or QGis will display it distorted (IrfanView without distortion).
> 
> Is there another way to get a non distorted image or what did I wrong?
> Martin
> 
> ..\FWTools2.4.2\bin>gdalinfo pic.tif
> Driver: GTiff/GeoTIFF
> Files: pic.tif
> Size is 15118, 11339
...
> gcp	mapX	mapY	pixelX	pixelY
> ------------------------------------------------
> upper left	27200	17200	0	0
> upper right	30400	17200	15118	0
> lower right	30400	14800	15118	-11339
> lower left	27200	14800	0	-11339
> 
> 
> Gcp	mapX	mapY	pixelX	pixelY
> -------------------------------------------------------
> upper left 13.4370322252183 52.4818154665202 0 0
> upper right 13.4841338099227 52.4818815506509 15118 0
> lower right 13.4842034283538 52.4603136448411 15118 -11339
> lower left 13.4371248633278 52.4602476117909 0 -11339

Martin,

Examining the GCPs in WGS84 space I see:

 >>> (13.4841338099227  - 13.4370322252183) / 15118
3.1155962894827404e-06
 >>> "%.10f" % ((13.4841338099227  - 13.4370322252183) / 15118)
'0.0000031156'
 >>> "%.10f" % ((52.4818154665202 - 52.4602476117909)/ 11339)
'0.0000019021'


Basically, in WGS84 space your source pixels are distinctly
non-square and gdalwarp distorts your image to produce square
pixels in the output projection.  If you want to preserve
the original aspect ratio try:

   -tr 0.0000031 0.0000019

on the gdalwarp commandline.

But basically WGS84 and Cassini-Soldner have quite different
characteristics in the area of your image, and WGS84 may not
be a good working coordinate system for the image.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent



More information about the gdal-dev mailing list