Opened 13 years ago

Last modified 13 years ago

#4086 closed defect

GCPs not read for Envisat MERIS reduced resolution scenes — at Version 4

Reported by: knutfrode Owned by: warmerdam
Priority: normal Milestone: 1.8.1
Component: GDAL_Raster Version: unspecified
Severity: normal Keywords: envisat, meris
Cc: antonio, knutfrode

Description (last modified by knutfrode)

Tested with GDAL versions 1.6 (Windows) and 1.8 (Linux), I find that GCP’s and coordinate systems (CS) are read for full resolution (MER_FR) scenes, but not for reduced resolution (MER_RR) scenes.

Here are the first output lines from gdalinfo for a reduced resolution (RR) scene:

Driver: ESAT/Envisat Image Format
Files: MER_RR__1PNRAL20100426_154828_000003662088_00498_42636_0001.N1
Size is 1121, 2081
Coordinate System is `'
Metadata:   
  MPH_PRODUCT=MER_RR__1PNRAL20100426_154828_000003662088_00498_42636_0001.N1
  MPH_PROC_STAGE=N
  MPH_REF_DOC=PO-RS-MDA-GS-2009_11_5A
  MPH_ACQUISITION_STATION=PDHS-K              
  ….

For corresponding full resolution scenes, CS and GCP’s are reported before the Metadata. This example scene, and others, can be downloaded from http://www.brockmann-consult.de/beam/data/products/MERIS

File structure for RR and FR scenes should be close to identical, so this probably requires a small fix only.


      

Change History (4)

comment:1 by antonio, 13 years ago

The method ScanForGCPs_MERIS relies on "LINES_PER_TIE_PT" and "SAMPLES_PER_TIE_PT" fields of the SPH, In MER_RR products such fields are not present.

Looking at http://earth.esa.int/pub/ESA_DOC/ENVISAT/Vol11_Meris_5b.pdf sect 11.4.1.7.3 values should be

LINES_PER_TIE_PT=16
SAMPLES_PER_TIE_PT=16

comment:2 by warmerdam, 13 years ago

I have downloaded MER_FRS_1PNRAL20100426_091349_000000822088_00494_42632_0001.N1 and a gdalinfo report on it does not include any GCPs, geotransform or coordinate system information. I was hoping to establish a working baseline to compare the _RR_ sets to.

comment:3 by warmerdam, 13 years ago

Component: defaultGDAL_Raster
Status: newassigned

In fact, I do see these lines in MER_RR1PNRAL20100426_154828_000003662088_00498_42636_0001.N1.

LINES_PER_TIE_PT=+016
SAMPLES_PER_TIE_PT=+016

Stepping through the code there seems to be a bunch of checking code for tie points per column based on the tiepoints per row logic. I have corrected this and now the MER_RR1PNR* dataset and the MER_FRS_1PN* dataset both have GCP sets.

The change is in trunk (r22403) and 1.8 (r22404).

Could someone do some additional testing of this?

Note, I also have a very small MER_RRBPNPDK20000620_103648_00000270A001_00180_00177_0642.N1 dataset that does not seem to have the LINES_PER_TIE_PT item or the corresponding tie point ADS. So apparently not all reduced resolution datasets will have GCPs.

comment:4 by knutfrode, 13 years ago

Description: modified (diff)

I have recompiled r22404, and can confirm that GCPs and coordinate systems are now read successfully for all the following tested scenes:

MER_FRS_1PNRAL20100426_091349_000000822088_00494_42632_0001.N1
MER_RR__2PNPDK20100519_101302_000005462089_00323_42962_4158.N1
MER_FR__2PNEPA20030714_102918_000000982018_00094_07162_0053.N1
MER_RR__2PNPDK20110515_100806_000004953102_00281_48140_3923.N1
MER_RR__1PNRAL20100426_154828_000003662088_00498_42636_0001.N1

Prior to the update it worked only for the MER_FR-scene. I believe the ticket can be safely closed. Thanks a lot!

Note: the archive of sample MERIS-files has in the meantime moved to: http://www.brockmann-consult.de/cms/web/beam/meris-products

Note: See TracTickets for help on using tickets.