[GRASSLIST:3320] Correcting SRTM with MrSID images

GRASS GRASS user
1

Hi all

I'm using r.mapcalc to get a raster layer to mask sea in SRTM data.

r.mapcalc red=r#image (good on bare rocky,beach or concrete coasts)
r.mapcalc green=g#image (good with vegetation)
r.mapcalc red2='if(red<50,null(),red)'
r.mapcalc green2='if(green<75,null(),green)'
r.mapcalc final='if(isnull(red2) && isnull(green2) && srtm<10,null(),srtm)'

This gives me a shoreline as good as the one I made by hand with
v.digit.There is still some speckle, but it can be eliminated with
little work.

Unfortunatelly I can't register SRTM and image data. After reprojecting
image from UTM-WGS84 to Lat_Lon-WGS84, it seems to be a shift of about
250-300 meters towards W in SRTM from image. More confusingly, this
shift is more evident in western shorelines than in eastern ones, where
often it seems even not having shift at all.¿?

Am I on LSD? Why can't I register the two datasets? How could I do? May
be this is a dumb question, but is driving me nuts. Any advice, pointer
or recommended reading would be greatly appreciated!

Please HELP!

Thanks a lot,

Roman

2

Hi Roman,

since you write that you have been using MrSID data
(did you use 'mrsiddecode'?), there is a probability
that either datum and/or ellipsoid are not matching.

From GRASS 5.3 onwards r.in.gdal is performing a test

to avoid that a user imports wrong data into a location.

You should verify your original data sets and compare
with the current location settings.

See also
http://mpa.itc.it/markus/shortcourse/notes3.html
how to use gdal_translate to assign new projection info
to a GeoTIFF file (I had to do that because 'mrsiddecode'
doesn't transfer this information when converting MrSID to
GeoTIFF).

Hope this helps

Markus

PS: The GSHHS vector shoreline data are not bad.

On Sat, May 01, 2004 at 05:53:07AM +0200, tatel@euskalnet.net wrote:

Hi all

I'm using r.mapcalc to get a raster layer to mask sea in SRTM data.

r.mapcalc red=r#image (good on bare rocky,beach or concrete coasts)
r.mapcalc green=g#image (good with vegetation)
r.mapcalc red2='if(red<50,null(),red)'
r.mapcalc green2='if(green<75,null(),green)'
r.mapcalc final='if(isnull(red2) && isnull(green2) && srtm<10,null(),srtm)'

This gives me a shoreline as good as the one I made by hand with
v.digit.There is still some speckle, but it can be eliminated with
little work.

Unfortunatelly I can't register SRTM and image data. After reprojecting
image from UTM-WGS84 to Lat_Lon-WGS84, it seems to be a shift of about
250-300 meters towards W in SRTM from image. More confusingly, this
shift is more evident in western shorelines than in eastern ones, where
often it seems even not having shift at all.¿?

Am I on LSD? Why can't I register the two datasets? How could I do? May
be this is a dumb question, but is driving me nuts. Any advice, pointer
or recommended reading would be greatly appreciated!

Please HELP!

Thanks a lot,

Roman

--
Markus Neteler <neteler itc it> http://mpa.itc.it
ITC-irst - Centro per la Ricerca Scientifica e Tecnologica
MPBA - Predictive Models for Biol. & Environ. Data Analysis
Via Sommarive, 18 - 38050 Povo (Trento), Italy