Opened 13 years ago

Closed 8 years ago

#1177 closed defect (invalid)

NITF and NITF with JPEG 2000 reading and writing errors

Reported by: larry@… Owned by: warmerdam
Priority: high Milestone:
Component: default Version: unspecified
Severity: major Keywords:
Cc:

Description (last modified by Even Rouault)

I'm not sure which of these issues are specific to the OSSIM NITF code and which
are going through the GDAL plugin:

I downloaded the Window binaries for OSSIM 1.6.4 and ran tests with iview and
imagelinker.

OSSIM is able to read large uncompressed and small JPEG 2000 files ok. It does
not seem to read large J2K files correctly.  It writes 3 flavors of NITF files
(nitf_block_band_separate, nitf_block_band_sequential and gdal_NITF). All three
of them have errors in the generated files.  There were structural and
compliancy (to Mil-Std-2500C) issues as well as cosmetic errors. In the
band_separate case a 3 and 7 band image were marked as MONO. In the
band_sequential and gdal case the file was marked as RGB or MULTI but the
IREPBAND fields were not consistent with the RGB marking. In that case OSSIM
used NITF 2.0 markings for a 2.1 file. (Note that even though the 2.0 standard
specifies the values OSSIM used the NITF community is currently using the 2.1
markings even for 2.0 files.) In the band_separate case the file length and
image length byte counts were off in the main header. In two of the cases the
files were marked as CLEVEL 1 even though CLEVEL 1 is not valid for an NITF 2.1
file.

A smart reader I tried was able to overcome the structural errors in the small
test cases and display the files. The raster data looked good.  In my large test
cases the raster data was corrupted as well. I read in a 401 MB 4 band file. In
one case it wrote back out only 349 MB and in other case it wrote over 11.7 GB
before I killed the process. In that case the Windows Task Manager said it had
written over 25 GB of data. The performance meter was still at 0% so who knows
where it would have gone.

You can email or call if you have more questions:
Larry Beck (404) 320-1824  larry@beckforce.com

Change History (1)

comment:1 Changed 8 years ago by Even Rouault

Description: modified (diff)
Resolution: invalid
Status: newclosed

Closing in the hope it is no longer valid

Note: See TracTickets for help on using tickets.