[Proj] GeoTiff <-> Proj.4

Gerald I. Evenden geraldi.evenden at gmail.com
Thu Oct 2 11:32:21 EDT 2008

On Thursday 02 October 2008 9:14:13 am support.mn at elisanet.fi wrote:
> Frank Warmerdam <warmerdam at pobox.com> kirjoitti:
> > The page at:
> >
> >   http://geotiff.maptools.org/proj_list/
> >
> > represents a "best effort" attempt to document the
> > correlation between geotiff definitions and proj.4
> > (as well as EPSG and OGC WKT).

I have already added some comments (probably unwelcome) about the contents of 
the above web site although I do issue kudos for the thought and effort 
behind it.

First, I detest the term "natural origin."  It seems to be synonymous with the 
basic mathematical origin.  "Natural" is an unnecessary adjective and just 
the term origin should suffice.

The projection origin can be "shifted" in either geographic or Cartesian space 
by respective use of lon_0 and sometimes lat_0 and false easting and northing 
(x_0 and y_0).  Use of lon_0 is ubiquitous among all types of projection 
usage and false easting/northing are universal among grid system 

lat_0 is sometimes ambiguous because it may actually specify the principle 
origin as in the case of many oblique projections.  But this may also be 
viewed as a simple shift like lon_0.

But use of lat_0 in normal cylindrical projections is exactly equivalent to 
using false northing.  In most grid system cases that I am aware of, false 
northing is used rather than lat_0.

Lastly, I cannot understand why it was necessary to include all the formulas 
for determining shift of the Cartesian coordinates when it is such a simple 
operation using [l]proj.  For example:

Mercator projection with origin for a map centered at CM 90w and with false 
origin located at 92W,33N.  First:

gie at charon:~$ lproj +proj=merc +ellps=WGS84 +lon_0=-90 -f '%.4f'
-92 33
-222638.9816    3872033.7329

>From the above numbers:

gie at charon:~$ lproj +proj=merc +ellps=WGS84 +lon_0=-90 \
     +x_0=222638.9816 +y_0=-3872033.7329
-92 33
0.00    -0.00
-90 35
222638.98       267339.03

The only thing "tricky" was to expand the output precision of the first pass 
to ensure that short precision would not show up in the cm precision of the 
final operations.  They still did with the -0.00 but ... .

Is that not simpler than the material in the mercator_2sp.html page?

When designing a grid system one only needs the easting and northing to be 
positive so the above numbers would probably be altered to:

> Yes, I have been reading that. There seem to be several
> problematic projecions. "Omerc" number 3, 4 and 5 in
> GeoTiff does not import to Proj4 at all, or does it?
> Geotiff number 6, "merc" is not used at all? Or should not
> be used?
> Geotiff number 15, "stere" uses "lat_0=CenterLat", not
> mentioned? Or is it only intergraph?
> Geotiff number 27, "tmerc" soth oriented can not be used
> directly, unless you turn all signs somewhere using extra
>  "ifs"? Isn't that forbidden, to have a south orientated map?
> How the people know where the north is if that is upside
> down?
> And Intergraph is using mostly different fields instead of
> those given in your link?
> ---------
> Linear names are easy: whatever given ->x_0 and y_0
> Ratio names also easy: whatever given -> k_0
> Angular names are most difficult and all are not using same
> names? There is no general rule? Ok. I make a mapping
> table for the names then.
> ----------
> BTW: There is couble of typos. Lcc(1) and Lcc(2) are entered
> both as number 9. I am assuming that Lcc(2) is number 8 and
> Lcc(1) is number 9?
> Regards: Janne.
> _______________________________________________
> Proj mailing list
> Proj at lists.maptools.org
> http://lists.maptools.org/mailman/listinfo/proj

The whole religious complexion of the modern world is due
to the absence from Jerusalem of a lunatic asylum.
-- Havelock Ellis (1859-1939) British psychologist

More information about the Proj mailing list