[GRASS-dev] Re: [GRASS GIS] #1088: r.fillnulls: support other interpolation methods

[GRASS GIS]

#1088: r.fillnulls: support other interpolation methods
-------------------------+--------------------------------------------------
 Reporter:  kyngchaos    |       Owner:  grass-dev@…              
     Type:  enhancement  |      Status:  new                      
 Priority:  normal       |   Milestone:  6.4.3                    
Component:  Raster       |     Version:  svn-develbranch6         
 Keywords:  r.fillnulls  |    Platform:  All                      
      Cpu:  All          |  
-------------------------+--------------------------------------------------
Changes (by hamish):

  * keywords:  fillnulls => r.fillnulls


Comment:

 > trialing bspline in grass7 now...

 the zoom trick for bspline is the same as in grass65: almost the same but
 slightly slower.

 I changed the bspline coeffs to si=3*res and lambda_i = 0.005. Should we
 change the default option answer to that? I note in the cross-validation
 code there the maximum they try is 0.05, so maybe 1.0 is way out of useful
 range.

 with that, time to complete was not quite as fast as zoomed-RST, but about
 twice as fast as in grass65. (partly due to si= settings, partly due to
 using some OpenMP, and perhaps partly due to speedups in trunk that MM was
 talking about?)

 > both RST and bspline results were (very nearly) identical to without the
 zoom.

 (there I meant r.univar results, but really that should be done on the
 cells-to-be patched in, not the final result where n_patched is quite a
 small percentage of the whole. an over-zoomed-in d.rast overlay didn't
 show discernible visual color change)


 Hamish

-- 
Ticket URL: <https://trac.osgeo.org/grass/ticket/1088#comment:13>
GRASS GIS <http://grass.osgeo.org>