[GRASSLIST:927] Re: interpolation & shading problem/question

[David Finlayson]

The rst algorithm breaks your data into processing cells to conserve 
memory. You need to tweak the parameters to minimize the artifact this 
sometimes creates. Dr. Mitsova (sp?) has several papers on this 
algorithm available on the internet.  If I recall correctly, she has a 
number of examples of settings on her web site too.

I use this algorithm a lot on very large datasets and you can get nice 
looking surfaces if you (a) have well spaced data (b) play with the 
settings to tune the surface to your data. I have seen the artifact 
myself when things were out of tune or the data was too sparse to make a 
nice surface.

Try googling for her and regular splines with tension for more information.

HTH

David

Brent Wood wrote:

 >
 >
 > I'm slowly getting to grips with the idiosyncrasies of GRASS, having used
 > other GIS & mapping tools for several years.
 >
 > My current problem lies in getting a good looking hillshade like the
 > spearfish dataset provides.
 >
 >
 > Hopefully I've included all the relevant info below....
 >
 >
 >
 > I have imported an xyz dataset into GRASS with:
 > cat mandes_.LI | awk '{print $2, $3, $4, $4}' | s.in.ascii andes d=3
 >
 > The file mandes_.LI comprises N X Y Z values, where N is a record ID that
 > I'm omitting when I build the site file. The Z value is depth in
 > meters, X & Y are long/latitude in decimal degrees.
 >
 >
 > The file imports to a site file as I expect, and I then interpolate it to
 > build a raster with
 >
 > s.surf.rst input=andes5m elev=andes5m tension=40 smooth=10 npmin=50
 >
 > This displays fine as a raster, showing the seabed topography I expect.
 >
 > I then try to get GRASS to display this with terrain (aspect) shading,
 > using the GRASS script below to create a new raster andes5m.shade
 >
 > shade.rel.sh altitude=20 azimuth=270 elevation=andes5m
 >
 >
 > This is then displayed using
 >
 > d.his h=andes5m i=andes5m.shade
 >
 > This picture is very grainy, the overall impression of shading is OK, but
 > the map appears to be built of bricks (about 32x32). I use the term 
bricks
 > because each has its own shading within its boundary, so pixel is not
 > appropriate, each brick comprises several pixels. I have tried changing
 > tension & smoothing parameters, without making much difference.
 >
 > The output below describes the elevation raster. If anyone can offer ay
 > advice to help resolve this, I would appreciate it.
 >
 >
 >
 > Thanks,
 >
 >   Brent Wood
 >
 >
 >
 >
 > 
+-----------------------------------------------------------------------+
 >  | Layer:    andes5m                     Date: Fri Aug  8 02:19:48 
2003  |
 >  | Mapset:   PERMANENT                      Login of Creator: baw 
     |
 >  | Location: andes 
     |
 >  | DataBase: /home/baw/grassdb 
     |
 >  | Title:     ( andes5m ) 
     |
 > 
|-----------------------------------------------------------------------|
 >  | 
     |
 >  |   Type of Map:  raster              Number of Categories: 255 
     |
 >  |   Data Type:    FCELL 
     |
 >  |   Rows:         1548 
     |
 >  |   Columns:      2268 
     |
 >  |   Total Cells:  3510864 
     |
 >  |        Projection: Latitude-Longitude (zone 0) 
     |
 >  |            N: 44:05:23.28S    S: 44:14:40.56S   Res: 0:00:00.36 
     |
 >  |            E: 174:22:53.04W    W: 174:36:29.52W   Res: 0:00:00.36 
     |
 >  |   Range of data:    min =  -1470.000000 max = 0.000000 
     |
 >  | 
     |
 >  |   Data Source: 
     |
 >  |    site file andes5m 
     |
 >  | 
     |
 >  | 
     |
 >  |   Data Description: 
     |
 >  |    generated by s.surf.rst 
     |
 >  | 
     |
 >  |   Comments: 
     |
 >  |    tension=40.000000, smoothing=10.000000 
     |
 >  |    dnorm=0.003973, dmin=0.000050, zmult=1.000000 
     |
 >  |    segmax=40, npmin=50, rmsdevi=3.464404 
     |
 >  |    zmin_data=-1470.000000, zmax_data=-485.000000 
     |
 >  |    zmin_int=-1470.000000, zmax_int=-489.320737 
     |
 >  |    @@t 
     |
 >  |    ]@(@@t 
     |
 >  |    ]@$ 
     |
 >  | 
     |
 > 
+-----------------------------------------------------------------------+
 >
 >
 >

-- 
David Finlayson
School of Oceanography
Box 357940
University of Washington
Seattle, WA  98195-7940
USA

Office: Marine Sciences Building, Room 112
Phone: (206) 616-9407
Web: http://students.washington.edu/dfinlays