Opened 16 years ago

Closed 12 years ago

#286 closed defect (fixed)

DTED -ve heights are wrong

Reported by: martin.daly@… Owned by: warmerdam
Priority: normal Milestone:
Component: GDAL_Raster Version: unspecified
Severity: normal Keywords:
Cc: neteler@…, martin.daly@…

Description (last modified by warmerdam)

Negative heights in DTED files are not calculated correctly.  An example is the 
lake (?) in the file w069_s50.dt0, sent separately by e-mail.

Attachments (1)

W06_S5.ZIP (265.3 KB) - added by martin.daly@… 16 years ago.
ZIP file containing DTED data

Download all attachments as: .zip

Change History (4)

Changed 16 years ago by martin.daly@…

Attachment: W06_S5.ZIP added

ZIP file containing DTED data

comment:1 Changed 16 years ago by warmerdam



The file in question appears to store the negative binary values in twos
complement form, yet section 3.11.1 of DTED spec MIL-PRF-89020A states:

"All elevation values are signed magnitude binary integers, right justified,
16 bits (2 bytes).  The sign bit is in the high order position.  Negative
values are not complemented."

While the code works fine if I treat the data as normal twos-complement
binary values, I can't help but wonder if this isn't just a problem with
a particular generating agency.  What was the source of the data you sent me?

I am going to search around a bit for other DTED data sources.  What I may
do is assume any value that comes out less than -16000 is indeed complemented
and adjust accordingly.  However, I am then concerned about how "nodata" areas
will be handled.  I don't seem to have any files with "nodata" areas but if
a marker like -32768 is used for these areas then my jiffy hack may screw up
for these values. 

Best regards,

comment:2 Changed 15 years ago by warmerdam

Martin Daly wrote:
> Frank,
> The data came from the GIS Data Depot, found by searching for DTED and
> Argentina:
> I had to register to be able to download it.  To stop you having to do
> the same I've attached the downloaded ZIP.


I tried out these images, and see they suffer from the same twos-complement
approach to negative values as the other image you provided.  The hack I
but in for #286 seems to work fine for them.  What I have yet to find 
is a DTED file with NULL pixels to ensure that I am handling them properly.

Best regards,

I set the clouds in motion - turn up   | Frank Warmerdam,
light and sound - activate the windows |
and watch the world go round - Rush    | Geospatial Programmer for Rent

comment:3 Changed 12 years ago by warmerdam

Component: defaultGDAL_Raster
Description: modified (diff)
Priority: highnormal
Resolution: fixed
Status: assignedclosed

I think this was just left open in the hopes of investigating null handling, but this many years later it seems unlikely one is going to turn up.


Note: See TracTickets for help on using tickets.