[gdal-dev] Thin Plate Spline Stability Solved?!?

Frank Warmerdam warmerdam at pobox.com
Sat Mar 29 23:50:05 EDT 2008


Folks,

I am working on some testing scripts for the transformers used for the warp
api and I encountered odd results from the thin plate spline transformer.

I dug into it, and somewhat miraculously tracked it to uninitialized
coefficient entries down in alg/thinplatespline.cpp.  I have now fixed
this in trunk, 1.5 branch and 1.4 branch.  I *suspect* this problem is
at the core of our long standing stability problems with the thin plate
spline algorithm.

I would appreciate it if folks who have encountered these problems in the
past could test out the fix, and make notes on what they find in:

    http://trac.osgeo.org/gdal/ticket/2300

If this works out, I'm so excited I think I'll push out the GDAL 1.4.5
release in the near future!  This whole thin plate spline problem has been
vexing me for several years.

PS. the bug was almost certainly introduced by me, when converting the
provided code to use VSIMalloc().

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    | President OSGeo, http://osgeo.org



More information about the gdal-dev mailing list