sources for altitude data (DEM)

Hello,

I've just noticed the optionnal elevation URL.
orux documented the "google API" way in the spanish manual:

http://maps.googleapis.com/maps/api/elevation/xml?locations=37,-2&sensor=true

FYI, "Geonames" is also providing such service:
http://ws.geonames.org/astergdemXML?lat=37&lng=-2
and
http://ws.geonames.org/astergdem?lat=37&lng=-2
with different syntax.

geonames elevation data is based on ASTER GDEM v2, which has a 30m precision at best (equator)
google API is based on SRTM which has a 90m precision but I'm not sure of that.

These 2 services apparently accepts only one coord, unlike mapquest.
It is sure useful for WPT and TRK altitude correction, but its use can be broader :
The NASA can provides elevation though WMS:
http://www.nasa.network.com/elev?REQUEST=GetCapabilities
especially its layer "mergedAsterElevations" which combine ASTER 30m ans USGS 10m
but unfortunatly, it doesn't answer to GetFeature requests, so the only output available is an image of format "image/bil" : this isn't useable for now in oruxmaps.

One great new feature would be to make a parralel query to this server each time oruxmaps update a tile, with the same parameters, BBOX and WIDTH/HEIGTH, in order to have accurate elevation data for the whole tile, and therefore the whole map, even for services that doesn-t provide DEM (almost all)
This way, the 3D view of oruxmaps would look awesome ! Topo + DEM =

jm

yip wrote:Hello,

I've just noticed the optionnal elevation URL.
orux documented the "google API" way in the spanish manual:

http://maps.googleapis.com/maps/api/elevation/xml?locations=37,-2&sensor=true

FYI, "Geonames" is also providing such service:
http://ws.geonames.org/astergdemXML?lat=37&lng=-2
and
http://ws.geonames.org/astergdem?lat=37&lng=-2
with different syntax.

geonames elevation data is based on ASTER GDEM v2, which has a 30m precision at best (equator)
google API is based on SRTM which has a 90m precision but I'm not sure of that.

These 2 services apparently accepts only one coord, unlike mapquest.
It is sure useful for WPT and TRK altitude correction, but its use can be broader :
The NASA can provides elevation though WMS:
http://www.nasa.network.com/elev?REQUEST=GetCapabilities
especially its layer "mergedAsterElevations" which combine ASTER 30m ans USGS 10m
but unfortunatly, it doesn't answer to GetFeature requests, so the only output available is an image of format "image/bil" : this isn't useable for now in oruxmaps.

One great new feature would be to make a parralel query to this server each time oruxmaps update a tile, with the same parameters, BBOX and WIDTH/HEIGTH, in order to have accurate elevation data for the whole tile, and therefore the whole map, even for services that doesn-t provide DEM (almost all)
This way, the 3D view of oruxmaps would look awesome ! Topo + DEM =

jm

Thanks, jm,

really valuable information; I will work in the future with it.

orux

Geonames expects long and lat values passed to it as separate variables, whereas mapquest and google expect a single long,lat variable. The mapquest/google format is what oruxmaps currently uses, which is incompatible with the geonames format.

Because of the different syntax , I have created my own php script to convert from what oruxmaps generates to what geonames requires... like an intermediate step from oruxmaps to geonames.

I have put the conversion script on a public server for general use, so try these settings in oruxmaps to access the geonames / aster gdem dataset:

* Coordinates per request: 20
* Url pattern: http://mikebrewer.co.uk/files/geonames_multiple.php?coord=${coord}&username=demo
* Coordinates separator: ,
* Xml elevation tag : astergdem

That's a single comma up there for the coordinates separator. For own use, you should create your own username rather than use the "demo" account above (replace "demo" with your own username)... you can request a free username here:

http://www.geonames.org/login

I'm really happy with the results... it gives a nice altitude profile that closely matches what I see in reality. As noted by the OP, the resolution of this dataset is good: 30m.

Hope this helps.

Mike