[Proj] Having all projection inverses was Re: Why

Robert B. Schmunk Robert.B.Schmunk at nasa.gov
Wed Oct 29 16:46:55 EDT 2008


Others have stated these points, but I'll reiterate...

In my own case, I have a couple applications which make map plots
on screen and which the user can then save to disk. Since the on-
screen map is a raster image, I need the inverse formulas for a
projection in order to determine from a pixel's coordinates the
corresponding lon-lat pair and therefore the appropriate data value
or color for that pixel. Further, users are able to click on a pixel
and have the application display a box with the lon-lat values and
the data values.

I must admit that I use neither PROJ or libproj for any of this,
and I'm only posting this msg because I check in on the PROJ mailing
list now and again because of a general interest in cartography. The
discussion about inverses and the Winkel Tripel caught my eye.

The code that I am using in the above mentioned applications is all in
Java and is all of my own writing. These applications allow for use
of any one of about 80 projections. Of course, many of the projections
are almost useless for practical purposes, but coding them up was a
fun academic exercise. Interestingly, though, of those 80 projections,
only five have so far required using the scheme described by Ipbuker
and Bildirici to get the inverse. These are the Winkel Tripel (of
course), two of Raisz's orthoapsidals, the Larrivee, and a Canters
polyconic. For a couple of these, picking the initial guesstimate
for the inverse was indeed a bit tricky; IIRC, Raisz's Armadillo was
the most troublesome.

Unfortunately, my code is not yet generally available for download.
I've been meaning to release it as open source for some time, but
have just not been a hurry to do the necessary organizational
paperwork. But anyone can feel free to e-mail me if you want to see
my code for a particular projection.

The applications I referred to above include the Panoply netCDF
viewer and the G.Projector projection explorer.
http://www.giss.nasa.gov/tools/panoply/
http://www.giss.nasa.gov/tools/gprojector/


--
Robert B. Schmunk
NASA Goddard Institute for Space Studies, 2880 Broadway, New York, NY  
10025



More information about the Proj mailing list