PDA

View Full Version : SRTM data - problems


BobH
08-11-2005, 01:40 PM
Has anybody found a solution to the problems of no data within the SRTM files. Either I have large mine shafts, if over 300m ASL or large blocks of flats under that !! I emailed the USGS about the problem over a year ago but never had a reply - I just wonder how others overcome the problems.

Pierre Thomas
08-12-2005, 06:42 PM
according to the SRTM project, scientists have decided not to populate the DEM with unreliable data. When the measurement system (radar interferometer) has detected wrong data (saturation from heavy refections on iced areas as example) or if retracing is not accurate between overlapped areas scanned, the data are removed and consequently the DEM file has "null data" that is presented in PL4 as 222 meter (on path profile) or unpopulated values in the terrain data file.
There are some freeware software that rebuild the missing areas by two principles:
- one solution is to use the GTOPO30 data for completing the DEM (but result is terrible, due to the high uncertainty of GTOPO30)
- a better one consist of calculation of the missing part by bilinear polynomial degree-n interpolation

Remember that the SRTM mission was effective in February 2000 and in certain countries, snow and iced areas drive to show DEM higher than the terrain level. Reversly, you may check that cover ground or constructions are generally not seen wit the effective height (between 10 to 20% in general)

Be also aware that if you use uncompleted DEM files, the "null data" areas are understood as 222m height for OHLOSS or diffraction analysis with PL4.

Regards,
Pierre

BobH
08-17-2005, 10:05 AM
Pierre,
do you know of a web site that has these programmes for "filling in the holes". I would like to try something, as I am working on a coastal area at the moment and the coverage predictions are way off due to these "null data" points.

Thanks

Jonnathan
08-17-2005, 08:31 PM
I 've found a tool. It is easy to use. It an Interpolate the null data or just change it for a especific value.

Null data is always showed below -32000 mts. So you can fix the ocean value (-10 to 3) selecting the range: -40000 to 2.

I have done and It works fine. I only check if I get null data in a mountain or hill (cut the hill to cero nd continue)

Rgds,

Jonnathan

Pierre THOMAS
08-24-2005, 11:15 AM
you will find intersting information at:
http://www.tec.army.mil/research/products/TD/tvd/survey/SRTMFill.html

Best regards,
Pierre

BobH
08-25-2005, 10:29 AM
Jonnathan,
Thanks for the info on Blackart. I have downloaded it and used it to great effect. The only problem that I find with the programme is that when trying to correct an area that is very close to sea level (<3m ASL) it is very difficult to get the programme to differentiate between sea and land.
Thanks again for the info.

BobH