[OSRS-PROJ] Re: Quick 'how-to' to on PROJ.4

Sun Jul 29 05:57:46 PDT 2001

Precedence: bulk
Reply-To: osrs-proj at remotesensing.org

Hi Petri and others,

i am tight on time, so i can not discuss this exhaustingly with you.

Only some keywords:

- datum shifts are only approximations, so there will always be an
error.
- You can not use the parameters for the 3 parameter datum shift with
the 7 parameter shift and vice versa. Two different sets of data.
- 7 parameter datum shift will give you better accuracy, but at the cost
of more calculation power needed. 
- there are different systems for the 7 parameter shift out which
disagree about the sign (+/-) of the parameters, so watch out that you
use the correct values with the right formula. 
- the parameters are calculated for specific regions, so check that the
region is correct. 

All this does not solve your problem, but should give you some things to
check. 15 arc seconds are IMHO much more than one could attribute to any
sort of accuracy error, this must be a systematic error. 

Another tip is that you should test with other systems (geotrans,
gpsman, gpstrans), so you can verify if this gives you values that are
in sync with the readings of your GPS device.

cu,

Andreas

"Petri J. Riipinen" wrote:
> 
> Precedence: bulk
> Reply-To: osrs-proj at remotesensing.org
> 
> Hello,
> 
> Hmm... I might be out of my league now, but do you think that the
> inaccuracy might come from the datum shift? I just learned that there are
> 3- and 7-parameter shifts. Should a 7-parameter shift be used for a better
> accuracy?
> 
> At least in this page:
> http://www.kms.dk/research/geodesy/reference/cotrans1_en.html there seems
> to be 7 parameters for WGS84 - ED50 - transformation. Would this help?
> 
> On this page: http://www.nls.fi/kartta/julkaisu/kkj.html there is some
> (IMO) simple transformation formula. Do you think that this formula would
> give me good enough results, if I can't get PROJ.4 to be accurate enough?
> At the end of the page there is some example that might be a good reference
> point when trying the transformations with cs2cs.
> 
> Regards,
> - Petri
> 
> At 17:32 27.7.2001 , Frank Warmerdam wrote:
> >"Petri J. Riipinen" wrote:
> > > Ok, I almost managed to do it, except that fl_transform returns an error
> > > code (100)
> > >
> > > First I want to mention that I found the Finnish KKJ-definitions in the
> > > epsg-file and they are below:
> > > # KKJ / Finland zone 1
> > > <2391> +proj=tmerc +lat_0=0.000000000 +lon_0=21.000000000 +k=1.000000
> > > +x_0=1500000.000 +y_0=0.000 +ellps=intl +units=m  no_defs <>
> > > # KKJ / Finland zone 2
> > > <2392> +proj=tmerc +lat_0=0.000000000 +lon_0=24.000000000 +k=1.000000
> > > +x_0=2500000.000 +y_0=0.000 +ellps=intl +units=m  no_defs <>
> > > # KKJ / Finland Uniform Coordinate System
> > > <2393> +proj=tmerc +lat_0=0.000000000 +lon_0=27.000000000 +k=1.000000
> > > +x_0=3500000.000 +y_0=0.000 +ellps=intl +units=m  no_defs <>
> > > # KKJ / Finland zone 4
> > > <2394> +proj=tmerc +lat_0=0.000000000 +lon_0=30.000000000 +k=1.000000
> > > +x_0=4500000.000 +y_0=0.000 +ellps=intl +units=m  no_defs <>
> > >
> > > I'm guessing that the Uniform Coordinate System is the one that I should
> > > use with Shapefiles and then different zones with different "vertical
> > > stripes" of Finland, when dealing with raster images.
> >...
> > > GPS-measurements:
> > > 1        6010.570       2455.930      GPS: FINLANDIA-HOUSE
> > > 2        6012.467       2456.127      GPS: RAILWAY YARD
> > > 3        6013.045       2458.826      GPS: VIIKKI MUSEUM
> > >
> > > Output in KKJ:
> > > 1        6674205.       2551915.      GPS: FINLANDIA-HOUSE
> > > 2        6677730.       2552047.      GPS: RAILWAY YARD
> > > 3        6678839.       2554525.      GPS: VIIKKI MUSEUM
> >
> >Petri,
> >
> >As suggested your main problem is that these angles need to be translated
> >into degrees minutse and seconds (or decimal degrees) rather than the odd
> >DDMM.mm format.
> >
> >Given that I tried a few experiments and still wasn't able to get very close
> >results.  The uniform coordinate system gave results that were way off, but
> >zone 2 gave me this (once without datum shift, once with).
> >
> >warmerda at gdal[139]% cs2cs +proj=tmerc +lat_0=0 +lon_0=24 +k=1.0
> >+x_0=2500000.0 +y_0=0.0 +ellps=intl +units=m +towgs84=-87,-98,-121 +to
> >+proj=latlong +ellps=intl
> >2551915 6674205
> >24d56'7.162"E   60d10'33.578"N 0.005
> >warmerda at gdal[140]% cs2cs +proj=tmerc +lat_0=0 +lon_0=24 +k=1.0
> >+x_0=2500000.0 +y_0=0.0 +ellps=intl +units=m +towgs84=-87,-98,-121 +to
> >+proj=latlong
> >+datum=WGS84
> >2551915 6674205
> >24d56'3.778"E   60d10'32.451"N 17.536
> >
> >However, this is still something like 15 seconds off in the easting which is
> >pretty dramatic.
> >
> >I am not sure what to do next to get proper results.
> >
> >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
> 
> <><><><><><><><>
>    Petri J. Riipinen
> petri.riipinen at nic.fi
> <><><><><><><><>
> ----------------------------------------
> PROJ.4 Discussion List
> See http://www.remotesensing.org/proj for subscription, unsubscription
> and other information.

-- 
Andreas Lange, 65187 Wiesbaden, Germany, Tel. +49 611 807850
Andreas.Lange at Rhein-Main.de - A.C.Lange at GMX.net
----------------------------------------
PROJ.4 Discussion List
See http://www.remotesensing.org/proj for subscription, unsubscription
and other information.