[Gvsig_english] raster iamge wrong band value

Benjamin Ducke benjamin.ducke at oxfordarch.co.uk
Wed Jul 7 10:34:29 CEST 2010


Just some more notes on this:

I tried changing the data type on that big GLOBCOVER file
that Klaus originally ran into trouble with. But SEXTANTE ran
out of memory. I then tried using the SEXTANTE map calculator
to add +128 to the data range, but same memory problems.

Luckily, the latest SEXTANTE 0.6 release comes with a GRASS GIS
interface. So I could use GRASS' r.out.gdal module to convert
to a new GeoTIFF of type "UInt16". With LZW compression enabled,
that gave me a GeoTIFF of pretty much exactly twice the size
of the original (which you would expect going from 8 bit Byte to 
16 bit Short).

Of course, one could also use the GDAL command line tools directly
to achieve the same effect.

But all in all, I must say that this would be a real obstacle
for many gvSIG users and I think this bug should really be fixed
for the 1.10 release, as it affects very basic functionality.
Like Klaus stated before, users should be able to rely on gvSIG
correctly loading basic file formats like GeoTIFF.

Given that a "proper" fix, changing all the data type handling
would probably take too much time, I am wondering if a quick fix
could consist in just "pretending" that a Byte type raster data
source is type Integer, and load it as integer instead of byte.

I would take a look at it but I am not sure where to start.
Which one is the method (in libRaster, I guess?) that first
initializes the data type for a raster data source?

Cheers,

Ben


----- Original Message -----
> Yes, short type is the right choice, of course.
> 
> About if it can be fixed or not, I don't know, I work mainly in
> vectorial part.
> 
> Cheers.
> 
> Benjamin Ducke escribió:
> > Or maybe use type "short"? Seems to me that would save some memory
> > (16 bit instead of 32 per cell).
> >
> > Is there any chance this fix will make it into 1.10?
> >
> > Cheers,
> >
> > Ben
> >
> > ----- Original Message -----
> >
> >> Hi Klaus.
> >>
> >> I think the problem is about byte datatype in java. In C++ you can
> >> use an unsigned byte, but in java you can't (always are signed).
> >>
> >> So, for this kind of images, we should use internaly int's instead
> >> of byte's. So, yes, it seems a bug, and can be corrected in future
> >> releases.
> >>
> >> Thanks for your help.
> >>
> >> Fran.
> >>
> >>
> >> Mithoefer, Klaus escribió:
> >>
> >>> Dear Guilermo,
> >>>
> >>>
> >>>
> >>> Thank you for your email. I have tried gvSIG, ESRI ArcGIS and
> >>> software which uses GDAL libraries such as FWtools. Its only gvSIg
> >>> which produces wrong results. The data type is byte(gvSIG) and
> >>> according to
> >>> ArcGIS 8 Bit unsigned integer which is to me generally the same
> >>> but more detailed information.
> >>>
> >>>
> >>>
> >>> However, I can off course work around that but to be honest it
> >>> seems it is a bug in gvSIG. If I can’t trust a software that it is
> >>> able to
> >>> read a file correctly (and it is a simple tif file public domain
> >>> land cover data.
> >>>
> >>>
> >>>
> >>> Thank you also for sending me the picture but it only showed the
> >>> landcover. Water bodies for example should have value 210.
> >>>
> >>>
> >>>
> >>> Again, thanks for your comments! That’s well appreciated.
> >>>
> >>>
> >>>
> >>> Klaus
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> *From:* gvsig_internacional-bounces at listserv.gva.es
> >>> [mailto:gvsig_internacional-bounces at listserv.gva.es] *On Behalf Of
> >>> *Guillermo Tamburini
> >>> *Sent:* Monday, July 05, 2010 09:30
> >>> *To:* Users and Developers mailing list
> >>> *Subject:* Re: [Gvsig_english] raster iamge wrong band value
> >>>
> >>>
> >>>
> >>> Sorry but,
> >>>
> >>> did you check my last proposal? Or was I wrong?
> >>>
> >>> This is an example of my result.
> >>>
> >>> Regards.
> >>>
> >>> Mithoefer, Klaus wrote:
> >>>
> >>> Thanks a lot Benjamin, but It didn't help.
> >>>
> >>> Kind regards,
> >>>
> >>> Klaus
> >>>
> >>> -----Original Message-----
> >>> From: gvsig_internacional-bounces at listserv.gva.es
> >>> <mailto:gvsig_internacional-bounces at listserv.gva.es>
> >>> [mailto:gvsig_internacional-bounces at listserv.gva.es] On Behalf Of
> >>> Benjamin Ducke
> >>> Sent: Thursday, July 01, 2010 12:54 AM
> >>> To: Users and Developers mailing list
> >>> Subject: Re: [Gvsig_english] raster iamge wrong band value
> >>>
> >>> Have you tried updating the layer's band statistics after
> >>> importing the raster?
> >>>
> >>> Ben
> >>>
> >>> ----- Original Message -----
> >>>
> >>>
> >>>     Thank you Guillermo and Vincent!
> >>>
> >>>
> >>>
> >>>     I guess Vincent might be right. The data source I am using can
> >>>     be
> >>>
> >>>     found on the web at
> >>>
> >>>     ftp://guestglobcover:Pir8Eefo@us-ext-nas.eo.esa.int/regional/07_Globcover_200412_200606_V2.2_Africa.zip
> >>>
> >>>     I use the file GLOBCOVER_200412_200606_V2.2_Africa_Reg.tif
> >>>     from this
> >>>
> >>>     package which is an 8 bit unsigned file format. It seems
> >>>     really that
> >>>
> >>>     gvSIG is interpreting this as 8 bit signed.
> >>>
> >>>
> >>>
> >>>     I tried to convert is using GDAL to Eras IMG, but the effect
> >>>     remains.
> >>>
> >>>
> >>>
> >>>     Thank you.
> >>>
> >>>
> >>>
> >>>     Klaus
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>     -----Original Message-----
> >>>
> >>>     From: gvsig_internacional-bounces at listserv.gva.es
> >>>     <mailto:gvsig_internacional-bounces at listserv.gva.es>
> >>>
> >>>     [mailto:gvsig_internacional-bounces at listserv.gva.es] On Behalf
> >>>     Of
> >>>
> >>>     Guillermo Tamburini
> >>>
> >>>     Sent: Wednesday, June 30, 2010 14:46
> >>>
> >>>     To: Users and Developers mailing list
> >>>
> >>>     Subject: Re: [Gvsig_english] raster iamge wrong band value
> >>>
> >>>
> >>>
> >>>     hi,
> >>>
> >>>
> >>>
> >>>     I have no idea what could be your problem. And i can not test
> >>>     with
> >>>
> >>>     other softwares because I haven't it now.
> >>>
> >>>
> >>>
> >>>     Well, my only proposal is that you try to make sure that the
> >>>     different
> >>>
> >>>     software are understanding the data in the same data type.
> >>>     Raster
> >>>
> >>>     layers could be byte, float, integer... Or to export the data
> >>>     to other
> >>>
> >>>     formats for testing if the result is always the same.
> >>>
> >>>
> >>>
> >>>     From gvSIG you could check the data type in the general tab of
> >>>     the
> >>>
> >>>     layer proprieties. May be in ESRI and in that case you get
> >>>     only real
> >>>
> >>>     positive numbers, and gvSIG is understanding numbers up and
> >>>     down the
> >>>
> >>>     zero. In that case you would obtain the data moved a half to
> >>>     the
> >>>
> >>>     negative values.
> >>>
> >>>
> >>>
> >>>     If you want you can send us some data example for testing it.
> >>>
> >>>
> >>>
> >>>     Best regards.
> >>>
> >>>
> >>>
> >>>     Mithoefer, Klaus wrote:
> >>>
> >>>
> >>>
> >>>         Dear List,
> >>>
> >>>
> >>>
> >>>         I use globecover vegetation data. When I try to retrieve
> >>>         the band
> >>>
> >>>         value by using the information picker or assign values to
> >>>         a point
> >>>
> >>>         layer I get the wrong value (e.g. for lakes I get-46
> >>>         instead of
> >>>
> >>>         210). I compared this with ArcGIS and FWtools.
> >>>
> >>>
> >>>
> >>>         I tried gvSIG 1.2, 1.9 and 2010 OADE with Geotif and Erdas
> >>>         IMG
> >>>
> >>>         format. They all do the same. What is the problem?
> >>>
> >>>
> >>>
> >>>         Thank you,
> >>>
> >>>
> >>>
> >>>         Klaus
> >>>
> >>>
> >>>
> >>>         Klaus Mithöfer
> >>>
> >>>         Geographic Information Systems (GIS) Officer
> >>>
> >>>         icipe - International Centre of Insect Physiology and
> >>>         Ecology
> >>>
> >>>         P.O. Box 30772 - 00100
> >>>
> >>>         Nairobi, Kenya
> >>>
> >>>         Phone: +254-20-8632000, Extn. 2147
> >>>
> >>>         Fax: +254-20-8632001/2
> >>>
> >>>         Mobile: +254-724-308404
> >>>
> >>>         E-Mail: kmithoefer at icipe.org <mailto:kmithoefer at icipe.org>
> >>>
> >>>         web: www.icipe.org <http://www.icipe.org>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>         ----------------------------------------------------------------------------
> >>>
> >>>         This position is supported by German development
> >>>         cooperation //
> >>>
> >>>         Centre for international Migration and Development -
> >>>
> >>>         http://www.cimonline.de/en/index.asp
> >>>
> >>>
> >>>
> >>>         Klaus.mithoefer at cimonline.de
> >>>         <mailto:Klaus.mithoefer at cimonline.de>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>         _______________________________________________
> >>>         Gvsig_internacional
> >>>
> >>>         mailing list
> >>>
> >>>         Gvsig_internacional at listserv.gva.es
> >>>         <mailto:Gvsig_internacional at listserv.gva.es>
> >>>
> >>>         http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>     -- Guillermo Tamburini Beliveau
> >>>
> >>>     Proyecto gvSIG
> >>>
> >>>     Consellería de Infraestructuras y Transportes (GVA)
> >>>
> >>>     Valencia (Spain)
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>     _______________________________________________
> >>>     Gvsig_internacional
> >>>
> >>>     mailing list
> >>>
> >>>     Gvsig_internacional at listserv.gva.es
> >>>     <mailto:Gvsig_internacional at listserv.gva.es>
> >>>
> >>>     http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
> >>>
> >>>     _______________________________________________
> >>>     Gvsig_internacional
> >>>
> >>>     mailing list
> >>>
> >>>     Gvsig_internacional at listserv.gva.es
> >>>     <mailto:Gvsig_internacional at listserv.gva.es>
> >>>
> >>>     http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> ------ Files attached to this email may be in ISO 26300 format
> >>> (OASIS Open Document Format). If you have difficulty opening them,
> >>> please visit http://iso26300.info for more information.
> >>>
> >>> _______________________________________________
> >>> Gvsig_internacional mailing list
> >>> Gvsig_internacional at listserv.gva.es
> >>> <mailto:Gvsig_internacional at listserv.gva.es>
> >>> http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
> >>> _______________________________________________
> >>> Gvsig_internacional mailing list
> >>> Gvsig_internacional at listserv.gva.es
> >>> <mailto:Gvsig_internacional at listserv.gva.es>
> >>> http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> -- Guillermo Tamburini Beliveau
> >>> Proyecto gvSIG
> >>> Consellería de Infraestructuras y Transportes (GVA)
> >>> Valencia (Spain)
> >>>
> >>> ------------------------------------------------------------------------
> >>>
> >>> _______________________________________________
> >>> Gvsig_internacional mailing list
> >>> Gvsig_internacional at listserv.gva.es
> >>> http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
> >>>
> >>>
> >> -- Fran Peñarrubia
> >> Scolab www.scolab.es
> >>
> >> Asociación gvSIG
> >> www.gvsig.com
> >>
> >> _______________________________________________ Gvsig_internacional
> >> mailing list
> >> Gvsig_internacional at listserv.gva.es
> >> http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
> >>
> >
> >
> > ------ Files attached to this email may be in ISO 26300 format
> > (OASIS Open Document Format). If you have difficulty opening them,
> > please visit http://iso26300.info for more information.
> >
> > _______________________________________________ Gvsig_internacional
> > mailing list
> > Gvsig_internacional at listserv.gva.es
> > http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional
> >
> 
> -- Fran Peñarrubia
> Scolab www.scolab.es
> 
> Asociación gvSIG
> www.gvsig.com
> 
> _______________________________________________ Gvsig_internacional
> mailing list
> Gvsig_internacional at listserv.gva.es
> http://listserv.gva.es/cgi-bin/mailman/listinfo/gvsig_internacional


------
Files attached to this email may be in ISO 26300 format (OASIS Open Document Format). If you have difficulty opening them, please visit http://iso26300.info for more information.



More information about the Gvsig_internacional mailing list