Problem with italian datum

[jcrepetto]

I have a problem with some datum and projections using the Rome prime Meridian with GDAL 1.5.0 .

Steps to reproduce the problem :

1. Download this sample map of Milan:

   wget http://www.cartografia.lispa.it/download/B6B2.zip
   

2. Extract the files

   unzip B6B2.zip
   

3. Convert the file to RGB :

   mv b6b2.tif b6b2.tif.orig
   pct2rgb.py b6b2.tif.orig b6b2.tif
   

4. Then convert the file to ECW.
   

There are two datum/projection pairs used by ErMapper :

• MONTEMAR/TMBOAGA1, based on the Greenwich prime meridian
  
• MONTROME/TMBOAG1R, based on the Rome prime meridian

gdal_translate -of ECW -co "TARGET=99" -co "DATUM=MONTROME"  -co "PROJ=TMBOAG1R"  -co "LARGE_OK=YES" b6b2.tif B6B2R.ecw 
gdal_translate -of ECW -co "TARGET=99" -co "DATUM=MONTEMAR"  -co "PROJ=TMBOAGA1"  -co "LARGE_OK=YES" b6b2.tif B6B2.ecw

5. Use gdalinfo to check the maps coordinates :

   gdalinfo B6B2.ecw
   Driver: ECW/ERMapper Compressed Wavelets
   Files: B6B2.ecw
   Size is 9449, 5906
   Coordinate System is:
   PROJCS["TMBOAGA1",
       GEOGCS["MONTE MARIO",
           DATUM["MONTEMAR",
               SPHEROID["INT24",6378388,297]],
           PRIMEM["Greenwich",0],
           UNIT["degree",0.0174532925199433]],
       PROJECTION["Transverse_Mercator"],
       PARAMETER["latitude_of_origin",0],
       PARAMETER["central_meridian",9.000000000000002],
       PARAMETER["scale_factor",0.9996],
       PARAMETER["false_easting",1500000],
       PARAMETER["false_northing",0],
       UNIT["Meter",1]]
   Origin = (1508000.000000000000000,5040000.413333333097398)
   Pixel Size = (0.846666666666670,-0.846666666666670)
   Corner Coordinates:
   Upper Left  ( 1508000.000, 5040000.413) (  9d 6'8.70"E, 45d30'45.50"N)
   Lower Left  ( 1508000.000, 5035000.000) (  9d 6'8.40"E, 45d28'3.47"N)
   Upper Right ( 1516000.153, 5040000.413) (  9d12'17.40"E, 45d30'45.01"N)
   Lower Right ( 1516000.153, 5035000.000) (  9d12'16.81"E, 45d28'2.98"N)
   Center      ( 1512000.077, 5037500.207) (  9d 9'12.83"E, 45d29'24.28"N)
   Band 1 Block=9449x1 Type=Byte, ColorInterp=Red
     Overviews: arbitrary
   Band 2 Block=9449x1 Type=Byte, ColorInterp=Green
     Overviews: arbitrary
   Band 3 Block=9449x1 Type=Byte, ColorInterp=Blue
     Overviews: arbitrary
   

   gdalinfo B6B2R.ecw
   Driver: ECW/ERMapper Compressed Wavelets
   Files: B6B2R.ecw
   Size is 9449, 5906
   Coordinate System is:
   PROJCS["TMBOAG1R",
       GEOGCS["MONTE MARIO",
           DATUM["MONTROME",
               SPHEROID["INT24",6378388,297]],
           PRIMEM["Greenwich",0],
           UNIT["degree",0.0174532925199433]],
       PROJECTION["Transverse_Mercator"],
       PARAMETER["latitude_of_origin",0],
       PARAMETER["central_meridian",-3.452333299999991],
       PARAMETER["scale_factor",0.9996],
       PARAMETER["false_easting",1500000],
       PARAMETER["false_northing",0],
       UNIT["Meter",1]]
   Origin = (1508000.000000000000000,5040000.413333333097398)
   Pixel Size = (0.846666666666670,-0.846666666666670)
   Corner Coordinates:
   Upper Left  ( 1508000.000, 5040000.413) (  3d20'59.70"W, 45d30'45.50"N)
   Lower Left  ( 1508000.000, 5035000.000) (  3d21'0.00"W, 45d28'3.47"N)
   Upper Right ( 1516000.153, 5040000.413) (  3d14'51.00"W, 45d30'45.01"N)
   Lower Right ( 1516000.153, 5035000.000) (  3d14'51.59"W, 45d28'2.98"N)
   Center      ( 1512000.077, 5037500.207) (  3d17'55.57"W, 45d29'24.28"N)
   Band 1 Block=9449x1 Type=Byte, ColorInterp=Red
     Overviews: arbitrary
   Band 2 Block=9449x1 Type=Byte, ColorInterp=Green
     Overviews: arbitrary
   Band 3 Block=9449x1 Type=Byte, ColorInterp=Blue
     Overviews: arbitrary
   

You can see that the longitudes of the second file are wrong.

[warmerdam]

This appears to represent a problem with the ecw_cs.dat coordinate system translation file. I'm not going to pursue it now though, I'm afraid.

[warmerdam]

In particular coordinate systems like MONTROME that should have a prime meridian other than greenwich do not. I'm going to contact Erdas to try and get more recent files and see if I can retranslate them.

[Jukka Rahkonen]

The file trunk/gdal/data/ecw_cs.wkt (GDAL 2.0) has still this info:

MONTROME,GEOGCS["MONTE MARIO",DATUM["MONTROME",SPHEROID["INT24",6378388,297]],PRIMEM["Greenwich",0],UNIT["degree",0.0174532925199433]]

[Jukka Rahkonen]

Still the same definition used in the ecw_cs.wkt file in 2017-11-22.

[Even Rouault]

In 41721:

data/ecw_cs.wkt: fix PRIMEM of MONTROME (fixes #2340)

[Even Rouault]

In 41722:

Revert unrelated changes accidentally committed with r41721 (refs #2340)