[Mapserver-dev] Advanced support of grid format (visualization, interrogation, analysis)

Frank Warmerdam warmerdam at pobox.com
Mon Oct 6 12:52:31 EDT 2003


Claude Philipona wrote:
> All web mapping solutions include the support of simple vector
> (visualisation, interrogation) and raster (visualisation) ? but few (or
> none) include the complete support for raster data (interrogation,
> analysis).  In desktop or professional GIS, these functions are included or
> in extensions (Spatial Analyst for Arcview, Vertical Mapper for MapInfo,
> etc.). The objective would be to extend Mapserver to support simple
> analysis function for raster (or grid) data (interrogation first,
> analysis). Mapserver uses GDAL lib that supports some grid formats
> (Arc/Info Binary Grid (.adf), Arc/Info ASCII Grid, Military Elevation Data
> (.dt0, .dt1), GRASS Rasters, USGS ASCII DEM (.dem), ?).
> The objective of the developpement would be to add to Mapserver
> the possibility to support visualisation, interrogation (and eventually
> simple analysis) for continuous data. The objective would not be to
> develop complex analysis functions available in desktop or professionnal
> GIS (at first) but to add some basic functions to present the results of
> these analysis. 
> Mapserver could return the pixel value for gridded data (elevation, slope,
> pollution concentration, average income,..). Simple function allowing to
> get information for a gridded format (height in the example below). We
> could use this function to build a more sophisticated function like a cross
> section.

Claude,

I have meant to kick in my 2 cents on this for a few days.  My short response
is that I don't think it makes sense to embed a huge amount of grid processing
directly in mapserver, but rather to do externally in environments like GRASS.

Another angle on it, is to use one or more of the mapscript environments to
do advanced processing.  For instance, GDAL has good bindings for Python
and numerical python.  This should make it possible to do some pretty
advanced stuff on grids in Numerical python and then visualize the result
with MapServer never even having to save the results to disk.

Where I think we do need to make advances in MapServer for "scientific"
raster data are:

1) Support CLASS operations on original raster values instead of only to the
    values computed after scaling to 8bit.  There are some performance issues
    with this that have held things up.

2) Support for "range based" coloring of raster data.  The ability to setup
    and apply color ramps to data, and to visualize them properly in the
    legend.

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    | Geospatial Programmer for Rent





More information about the mapserver-dev mailing list