Re: [OSSIM] Weird behaviour when reading tiff files (geographic referencing)

[Garrett P. <gar...@gm...>]

Hello Dave:

Cool!  Will look at it and add a bug ticket


Thank you

Garrett


> On Sep 21, 2018, at 9:03 AM, David Burken <db...@gm...> wrote:
> 
> Hi Roberta,
> 
> Good catch...
> 
> Garrett, Roberta is correct, you're not picking up the raster type, i.e. point, area and you're off by 1/2 pixel.
> 
> $ ossim-info -d srtm_47_04.tif
> tiff.image0.raster_type:  pixel_is_area
> tiff.image0.model_pixel_scale:  """0.000833333333333333 0.000833333333333333 0 """
> tiff.image0.model_tie_point:  """0 0 0 49.999583817611 45.0004168845861 0 """
> 
> // Bad:
> < image0.geometry.tie_point_lat:  45.0004168845861
> < image0.geometry.tie_point_lon:  49.999583817611
> ---
> // Good:
> > image0.geometry.tie_point_lat:  45.0000002179194
> > image0.geometry.tie_point_lon:  50.0000004842777
> 
> You can download some samples at:
> 
> http://srtm.csi.cgiar.org/SELECTION/listImages.asp <http://srtm.csi.cgiar.org/SELECTION/listImages.asp>
> 
> Take care,
> Dave
> 
> 
> 
> On 9/20/18 4:34 PM, Roberta Ravanelli wrote:
>> Hi all, 
>> 
>> I would like to update my Ossim installation from an old version installed in an old pc (Ubuntu 12.04, libgeotiff 2, opencv 2) to a newer version in a more recent pc (Ubuntu 16.04, libgeotiff 5, opencv 4).
>> 
>> I've actually managed to compile and run Ossim in the newer pc, but I've noticed a different behaviour between the two versions when reading tiff files.
>> Specifically, if the tiff has a geographic georeferencing (lon/lat), there is a wrong shift of a half pixel in the coordinate of the upper left coordinate (the tiepoint, as it is called in the ossimImageGeometry object).
>> 
>> Below an example on a SRTM v4_1 tile (srtm_39_03.tif):
>> 
>> OLDER VERSION (right: UL as in the .tfw file)
>> type: ossimImageGeometry
>>   No transform defined. Using identity transform.
>>   m_projection: 
>> // ossimMapProjection::print
>> type:  ossimEquDistCylProjection
>> major_axis:  6378137.000000000000000
>> minor_axis:  6356752.314199999906123
>> origin_latitude:  0.000000000000000
>> central_meridian:  0.000000000000000
>> origin: ( 0.000000000000000, 0.000000000000000, 0.000, WGE )
>> datum:  WGE
>> meters_per_pixel_x:  92.6625433148566
>> meters_per_pixel_y:  92.6625433148566
>> false_easting_northing: (0,0)
>> false_easting_northing_units: meters
>> pcs_code: 4326
>> tie_point_xy: (10.0000002905424,50.0000003631855)
>> tie_point_units: degrees
>> pixel_scale_xy: (0.000833333333333333,0.000833333333333333)
>> pixel_scale_units: degrees
>> ossimErrorStatusInterface::print
>> theErrorStatus:         0
>> theErrorStatus string:  OSSIM_OK
>> m_decimationFactors[0]: ( 1.000000000000000, 1.000000000000000 )
>> m_imageSize: ( 6001, 6001 )
>> m_targetRrds: 0
>> 
>> 
>> Newer version (WRONG: UL as in the .hdr file)
>> Type: ossimImageGeometry
>>   No transform defined. Using identity transform.
>>   m_projection: 
>> // ossimMapProjection::print
>> type:  ossimEquDistCylProjection
>> major_axis:  6378137.000000000000000
>> minor_axis:  6356752.314199999906123
>> origin_latitude:  0.000000000000000
>> central_meridian:  0.000000000000000
>> origin: ( 0.000000000000000, 0.000000000000000, 0.000, WGE )
>> datum:  WGE
>> meters_per_pixel_x:  92.6625433148566
>> meters_per_pixel_y:  92.6625433148566
>> false_easting_northing: (0,0)
>> false_easting_northing_units: meters
>> pcs_code: 4326
>> tie_point_xy: (9.99958362387572,50.0004170298522)
>> tie_point_units: degrees
>> pixel_scale_xy: (0.000833333333333333,0.000833333333333333)
>> pixel_scale_units: degrees
>> ossimErrorStatusInterface::print
>> theErrorStatus:         0
>> theErrorStatus string:  OSSIM_OK
>> m_decimationFactors[0]: ( 1.000000000000000, 1.000000000000000 )
>> m_imageSize: ( 6001, 6001 )
>> m_targetRrds: 0
>> 
>> 
>> I would like to know if such a different behaviour may depend on the changes performed in Ossim (for instance, I've noticed that the ossimGeoTiff class is changed a lot beetween the two verions) or maybe it is caused by the changes in the dependencies (maybe libgeotiff?)
>> 
>> Can you give me a hint, since I have to abandon the oldest pc?
>> 
>> Thank you very much!
>> 
>> Roberta
>> 
>> --------------------------------------------------------------------------------
>> Roberta Ravanelli, PhD 
>> Geodesy and Geomatics Division   
>> University of Rome "La Sapienza"
>> Via Eudossiana, 18 - 00184 Rome Italy
>> E-mail rob...@un... <mailto:rob...@un...>
>> 
>> 
>> 
>> _______________________________________________
>> www.ossim.org <http://www.ossim.org/>
>> Ossim-developer mailing list
>> Oss...@li... <mailto:Oss...@li...>
>> https://lists.sourceforge.net/lists/listinfo/ossim-developer <https://lists.sourceforge.net/lists/listinfo/ossim-developer>
> 
> _______________________________________________
> www.ossim.org
> Ossim-developer mailing list
> Oss...@li...
> https://lists.sourceforge.net/lists/listinfo/ossim-developer