| What is SRTM?
For many years we at Geomantics have been frustrated by the
lack of good-quality low-cost heightfield data for landscapes
outside the USA. European and Asian data in particular has
traditionally been expensive to obtain. This has now changed
with release by NASA of full Euroasian data from the Shuttle
Radar Topography Mission into the public domain - a 3-arc
second dataset (an approximatly 90 metre grid at the equator)
significantly better than anything previously available.
However the SRTM data so far released is in a raw format containing
small undefined patches, poorly defined coastlines and other
artifacts. In particular the data is released in latitude/longitude
cell format, so that if used directly in many applications
it appears grossly distorted.
What processing has Geomantics
applied?
We have taken the raw data and applied the following editing
passes :-
- Missing gridpoints defined
using nngridr interpolation. We believe that this is the
best interpolation method currently available
- Sealevels set to zero and
coastline defined. This is a significant problem with the
raw SRTM data as tides and other artifacts significantly
distort the sea surface (sealevel heights range from -5
to +5 meters) and hence the coastline. For the UK data we
have manually checked processing results for errors.
- Geocorrect the data carry
out any necessary interpolation for the coordinate system
selected. Most visualization, modelling and mapping programs
expect that heightfields will be a regular square grid.
Because the SRTM data is in spherical coordinates this is
very much not the case at European latitudes. UK data for
example has been transformed to UK National Grid format.
- Generate tiles.
Note we have not applied any
processing to freshwater bodies. These appear a little uneavenly
but the height variation is not as significant as that for
sea water. We may apply further processing to these areas
and if so the enhanced data will be available as a free upgrade.
But isn't NIMA going to clean
up the data anyway?
According to NASA the data will be cleaned up by NIMA and
issued to the public domain in processed form. However no
timescale is given for this and it won't be available in corrected
format - i.e. National Grid for the British data. When the
cleaned data is available we will process it and make it available
as a free upgrade.
SRTM data is free - so how
can you justify charging?
We are not charging for the original data - you can happily
download that here.
Our charges simply reflect our efforts in processing the data
and making it available in a user-friendly form.
In what formats can you supply
the data?
For simplicity we are supplying the data in 16bit raw format
as default - there are many packages that can handle this
including our own Landformer Pro. For Terragen and Bryce users
we also supply a utility to read in 4 adjacent raw files and
select any area from these into a .ter or .pgm file.
Generally we are resampling the
data to a 100 metre grid and supplying in 250 x 250 grid tiles.
The original SRTM is supplied as 3 arc seconds, so approximatly
90 metre by 50 metre at European latitude.
It is possible to resample from
our database at other resolutions - for example for Great
Britain we can supply data sampled to 50 metre in OS Panorama
NTF format. We can also supply larger tiles centred on particular
locations - please contact us for details.
What are the limitations on
the data?
The SRTM data is by far the most cost effective high resolution
dataset so far available - for Great Britain for example we
are charging approximatly 2% of the cost of analogous data
from the Ordnance Survey. However the quality of the data
is not as good as that typically obtained from National Mapping
agencies, so while the data is more than adequate for generating
just about any form of visualization or preliminary GIS analysis,
we wouldn't suggest you base the final decision about your
mobile phone mast positioning on it.
Any licence conditions?
Our processed data is sold on a single user licence basis.
Please contact us for multi-user prices. Our processed data
may not be redistributed.
|
