[Geoserver-users] Projection problem of a gdalwarp EPSG:900913 file

Dear List,

This is geoserver 2.0.2.

I use geoserver to display Landsat7 scenes that were downloaded from the
GLCF website. The files came with the EPSG:32635 projection. I imported the
scenes in grass, pan-sharpened the image (with i.fusion.brovey), formed a
true-color indexed output image (with r.composite) and output it to a
single-band, 16 bits GeoTiFF (with r.out.gdal). I then used gdal_translate
to transform it into a 8-bits rgba GeoTiFF image :

gdal_translate -expand rgba -ot Byte -a_nodata 0 00_marm_pan.rgb.tif
00_marm_pan.rgb.tif

I then used gdalwarp to mosaic and reproject two adjacent files into on
Google Merkator projection :

gdalwarp -co "TILED=YES" -t_srs EPSG:900913 00_marm_pan.rgb_900913.tif
00_bb_pan.rgb_900913.tif Landsat_2000.tif

I then add overviews with :
gdaladdo Landsat_2000.tif 2 4 8 16"

I tried to display the resulting file in grass, qgis and openjump. I all
these software, my file overlays nicely with all other vector layers of same
projection and google tiles (qgis with openlayers plugin, see the first
screenshot). I then imported the file in geoserver but here the image is far
off of the same vector layers tested during the previous stage. See the
second screenshot.

The resulting mosaic file reports :
Size is 1507, 920
Coordinate System is:
PROJCS["Google Maps Global Mercator",
    GEOGCS["WGS 84",
        DATUM["WGS_1984",
            SPHEROID["WGS 84",6378137,298.257223563,
                AUTHORITY["EPSG","7030"]],
            AUTHORITY["EPSG","6326"]],
        PRIMEM["Greenwich",0],
        UNIT["degree",0.0174532925199433],
        AUTHORITY["EPSG","4326"]],
    PROJECTION["Mercator_1SP"],
    PARAMETER["central_meridian",0],
    PARAMETER["scale_factor",1],
    PARAMETER["false_easting",0],
    PARAMETER["false_northing",0],
    UNIT["metre",1,
        AUTHORITY["EPSG","9001"]]]
Origin = (3203626.613789663650095,5031755.790616016834974)
Pixel Size = (37.731260645376345,-37.731260645376345)
Metadata:
  AREA_OR_POINT=Area
Image Structure Metadata:
  INTERLEAVE=PIXEL
Corner Coordinates:
Upper Left ( 3203626.614, 5031755.791) ( 28d46'43.20"E, 41d19'19.76"N)
Lower Left ( 3203626.614, 4997043.031) ( 28d46'43.20"E, 41d 5'11.94"N)
Upper Right ( 3260487.624, 5031755.791) ( 29d17'22.05"E, 41d19'19.76"N)
Lower Right ( 3260487.624, 4997043.031) ( 29d17'22.05"E, 41d 5'11.94"N)
Center ( 3232057.119, 5014399.411) ( 29d 2'2.63"E, 41d12'16.23"N)
Band 1 Block=1507x1 Type=Byte, ColorInterp=Red
  Overviews: 754x460, 377x230, 189x115, 95x58
  Mask Flags: PER_DATASET ALPHA
  Overviews of mask band: 754x460, 377x230, 189x115, 95x58
Band 2 Block=1507x1 Type=Byte, ColorInterp=Green
  Overviews: 754x460, 377x230, 189x115, 95x58
  Mask Flags: PER_DATASET ALPHA
  Overviews of mask band: 754x460, 377x230, 189x115, 95x58
Band 3 Block=1507x1 Type=Byte, ColorInterp=Blue
  Overviews: 754x460, 377x230, 189x115, 95x58
  Mask Flags: PER_DATASET ALPHA
  Overviews of mask band: 754x460, 377x230, 189x115, 95x58
Band 4 Block=1507x1 Type=Byte, ColorInterp=Alpha
  Overviews: 754x460, 377x230, 189x115, 95x58

Geoserver reports the native SRS of this file as UNKNOWN. There is a blue
link next to it called "unnamed" , when I click on it, I get :
PROJCS["unnamed",
  GEOGCS["WGS 84",
    DATUM["World Geodetic System 1984",
      SPHEROID["WGS 84", 6378137.0, 298.257223563,
AUTHORITY["EPSG","7030"]],
      AUTHORITY["EPSG","6326"]],
    PRIMEM["Greenwich", 0.0, AUTHORITY["EPSG","8901"]],
    UNIT["degree", 0.017453292519943295],
    AXIS["Geodetic longitude", EAST],
    AXIS["Geodetic latitude", NORTH],
    AUTHORITY["EPSG","4326"]],
  PROJECTION["Mercator_1SP"],
  PARAMETER["latitude_of_origin", 0.0],
  PARAMETER["central_meridian", 0.0],
  PARAMETER["scale_factor", 1.0],
  PARAMETER["false_easting", 0.0],
  PARAMETER["false_northing", 0.0],
  UNIT["m", 1.0],
  AXIS["Easting", EAST],
  AXIS["Northing", NORTH]]

The Geotif file just before reprojection reports :
Size is 1487, 885
Coordinate System is:
PROJCS["WGS 84 / UTM zone 35N",
    GEOGCS["WGS 84",
        DATUM["WGS_1984",
            SPHEROID["WGS 84",6378137,298.257223563,
                AUTHORITY["EPSG","7030"]],
            AUTHORITY["EPSG","6326"]],
        PRIMEM["Greenwich",0],
        UNIT["degree",0.0174532925199433],
        AUTHORITY["EPSG","4326"]],
    PROJECTION["Transverse_Mercator"],
    PARAMETER["latitude_of_origin",0],
    PARAMETER["central_meridian",27],
    PARAMETER["scale_factor",0.9996],
    PARAMETER["false_easting",500000],
    PARAMETER["false_northing",0],
    UNIT["metre",1,
        AUTHORITY["EPSG","9001"]],
    AUTHORITY["EPSG","32635"]]
Origin = (649806.344271000009030,4554888.908224999904633)
Pixel Size = (28.506380857431072,-28.402912233898281)
Metadata:
  AREA_OR_POINT=Area
Image Structure Metadata:
  INTERLEAVE=PIXEL
Corner Coordinates:
Upper Left ( 649806.344, 4554888.908) ( 28d47'5.17"E, 41d 7'53.36"N)
Lower Left ( 649806.344, 4529752.331) ( 28d46'43.20"E, 40d54'18.59"N)
Upper Right ( 692195.333, 4554888.908) ( 29d17'21.99"E, 41d 7'21.22"N)
Lower Right ( 692195.333, 4529752.331) ( 29d16'53.82"E, 40d53'46.71"N)
Center ( 671000.838, 4542320.620) ( 29d 2'1.11"E, 41d 0'50.97"N)
Band 1 Block=1487x1 Type=Byte, ColorInterp=Red
  NoData Value=0
Band 2 Block=1487x1 Type=Byte, ColorInterp=Green
  NoData Value=0
Band 3 Block=1487x1 Type=Byte, ColorInterp=Blue
  NoData Value=0
Band 4 Block=1487x1 Type=Byte, ColorInterp=Alpha
  NoData Value=0

I can't help but blame geoserver for this as 3 other programs displayed it
correctly. Anyone with a clue?

million thanks beforehand
servet

http://old.nabble.com/file/p29562707/scnsht1.jpeg
http://old.nabble.com/file/p29562707/scnsht2.png
--
View this message in context: http://old.nabble.com/Projection-problem-of-a-gdalwarp-EPSG%3A900913-file-tp29562707p29562707.html
Sent from the GeoServer - User mailing list archive at Nabble.com.

I know many of you people are busy for the approaching FOSS4G but I
desperately need help in this and your valuable comments would be greatly
appreciated...

I made the resulting file available for download just in case someone is
willing to help me :
http://ekumen.homelinux.net/Landsat_2000_EPSG_900913.tif

please let me know if I can provide more diagnostic info
thanks from advance
Servet
--
View this message in context: http://old.nabble.com/Projection-problem-of-a-gdalwarp-EPSG%3A900913-file-tp29562707p29611749.html
Sent from the GeoServer - User mailing list archive at Nabble.com.

On Fri, Sep 3, 2010 at 3:18 AM, midyetava <sa.cizmeli@anonymised.com> wrote:

I know many of you people are busy for the approaching FOSS4G but I
desperately need help in this and your valuable comments would be greatly
appreciated...

It's not clear but what did you tell GeoServer the projection for your
image was?

Ian
--
Ian Turton

Ian Turton wrote:

It's not clear but what did you tell GeoServer the projection for your
image was?

I set the declared SRS to 900913, the same target SRS as I used in with
gdalwarp. Geoserver does not recognize it (UNKNOWN) however it is able to
show some of the projection parameters in the blue link next to the UNKNOWN
tag (as given in my first message).

I am not able to play with the SRS handling parameter. No matter what I set,
geoserver automatically sets it back to "Reproject Native to Declared".
Clearly geoserver has understanding recognizing projection information other
GIS can recognize.

A link to the image in question was posted in my previous message. Please
let me know if I can be of more help.

Servet
--
View this message in context: http://old.nabble.com/Projection-problem-of-a-gdalwarp-EPSG%3A900913-file-tp29562707p29616843.html
Sent from the GeoServer - User mailing list archive at Nabble.com.

midyetava wrote:

...
I tried to display the resulting file in grass, qgis and openjump. I all
these software, my file overlays nicely with all other vector layers of
same projection and google tiles (qgis with openlayers plugin, see the
first screenshot). I then imported the file in geoserver but here the
image is far off of the same vector layers tested during the previous
stage.
...
I can't help but blame geoserver for this as 3 other programs displayed it
correctly. Anyone with a clue?

Yes this is a problem with geoserver's 900913 WKT definition not matching
the GDAL WKT in the image header, hence it doesn't recognise and and tries
incorrectly to reproject the raster. See
http://old.nabble.com/geoserver-problems-with-gdal's-900913-projection-definition-td30648435.html
http://old.nabble.com/geoserver-problems-with-gdal's-900913-projection-definition-td30648435.html
for a workaround.
--
View this message in context: http://old.nabble.com/Projection-problem-of-a-gdalwarp-EPSG%3A900913-file-tp29562707p30653318.html
Sent from the GeoServer - User mailing list archive at Nabble.com.