Ticket #2105 (closed defect: fixed)

Opened 8 months ago

Last modified 8 months ago

RADARSAT-2 dataset does not expose a subdataset for uncalibrated DNs.

Reported by: pvachon Assigned to: pvachon
Priority: high Milestone: 1.5.1
Component: GDAL_Raster Version: unspecified
Severity: normal Keywords: radarsat2, rs2
Cc: allanx13

Description

The RADARSAT-2 driver does not expose a subdataset in the subdataset metadata domain that is representative of the uncalibrated digital numbers. While this isn't an issue if you read the file without regard for the subdatasets domain, applications like OpenEV do not provide the option to not open a subdataset item, which means that the uncalibrated data cannot be opened for viewing.

Change History

12/18/07 19:58:59 changed by pvachon

  • status changed from new to closed.
  • resolution set to fixed.

Fixed in r13373 in TRUNK. I will ask Frank if he minds applying these fixes to the 1.5.0 release candidate. Otherwise, we'll have to wait until 1.5.1 (but if you're working against trunk, it probably is a non-issue).

(follow-up: ↓ 3 ) 12/19/07 09:41:08 changed by allanx13

The ScanSAR Wide A image that I had mentioned in ticket 2104 now opens and displays properly in OpenEV, whether I choose to display the uncalibrated DNs, sigma nought, beta, or gamma.

On the other hand, I also have a Standard 1 16-bit product. I can only display the uncalibrated DNs of this image in OpenEV; if I pick sigma nought, beta, or gamma, I get a ton of IReadBlock and GetBlockRef? failure messages.

This may end up not being a problem as we're still just dealing with simulated products while waiting for the real products to start coming in. Perhaps there's less need to spend more time on this driver before that happens?

I should also mention that I've been using the trunk version of rs2dataset.cpp in an otherwise standard copy of version 1.4.4.

(in reply to: ↑ 2 ) 12/19/07 10:01:05 changed by pvachon

  • status changed from closed to reopened.
  • version changed from 1.5.0 betas/RCs to unspecified.
  • resolution deleted.
  • milestone changed from 1.5.0 to 1.5.1.

Allan,

Are you able to provide me with a copy of the offending product? I've only been able to test SLC Polarimetric products; that is all I have access to. Otherwise, if you could provide me with a log of the output from gdal_translate for that dataset (maybe translating it to GeoTIFF), that would also be quite helpful.

Regards, Phil

Replying to allanx13:

The ScanSAR Wide A image that I had mentioned in ticket 2104 now opens and displays properly in OpenEV, whether I choose to display the uncalibrated DNs, sigma nought, beta, or gamma. On the other hand, I also have a Standard 1 16-bit product. I can only display the uncalibrated DNs of this image in OpenEV; if I pick sigma nought, beta, or gamma, I get a ton of IReadBlock and GetBlockRef? failure messages. This may end up not being a problem as we're still just dealing with simulated products while waiting for the real products to start coming in. Perhaps there's less need to spend more time on this driver before that happens? I should also mention that I've been using the trunk version of rs2dataset.cpp in an otherwise standard copy of version 1.4.4.

12/19/07 11:13:08 changed by allanx13

Hi Phil,

I can provide you a copy of the Standard 1 product as well as the ScanSAR product that works. Can you send me an e-mail at allan.meek@drdc-rddc.gc.ca so I can provide the directions for FTPing the file?

Thanks, Allan

12/19/07 14:01:36 changed by pvachon

  • status changed from reopened to closed.
  • resolution set to fixed.

Allan,

Give it a try now. It worked for me using the products you provided. The revision of interest is r13399.

Regards, Phil

12/20/07 09:18:15 changed by allanx13

Hi Phil,

I can now display the uncalibrated DNs, the sigma nought, the beta, and the gamma for both the Standard 1 product and the ScanSAR Wide A product in OpenEV. Thanks for the fix!

Allan