Earthdata Forum

Currently, I try to process some MERIS data during 2002-2012 by the SeaDAS 7.3. The L2_FRS_OC product generated by l2gen is ok , but when I used the gpt.sh to make mosaic with reprojection and subset, the result is very weird. The width and height of mapped data is just 20*15 for a 2000km2 region(30.8,31.8,119.8,120.8).
The command is : gpt.sh /home/rslakes/optics/tmp/M20070106_gpt.xml -t ./xxx.h5 -f 'HDF5' /home/rslakes/optics/tmp/M20070106_QKM.L2_FRS_OC,
the xml file for mosaic and a composite png will be putted in attachments.
I don't know why this case occured? The similar script works well for other data, and I can't find the errors in xml file.

Could someone give me some suggestions?

Regards,
zhigang
attachment 1

attachment 2

Hi Zhigang,

Please make your input files available for us and we will take a look. Meanwhile, please use the mosaic tool through GUI and see if the results are the same.

Thanks,
Aynur

Thanks Aynur. The result generated by mosaic tool though GUI is OK. Actually, this problem is common for those data after 2006, yet the results for those data in 2002-2005 are OK, I don't know why.
Due to the input file size is large, I will send the data to your e-mail box (aynur.abdurazik@nasa.gov).
Looking forward to your reply.

Happy Chinese New Year!
Zhigang

Dear Aynur,
The input file size is ~48M, which exceeds sever nasa.gov(xx.xxx.xx.xx) limit, the mail was rejected by server.
And, how can I send the input file to you ?

Regards,
zhigang

Have you tried adding the pixel size variables, e.g., from gpt.command Mosaic -h

  -PpixelSizeX=<double>            Size of a pixel in X-direction in map units.
                                   Default value is '0.05'.
  -PpixelSizeY=<double>            Size of a pixel in Y-direction in map units.
                                   Default value is '0.05'.

With units of degrees, the above default gives 20 pixels per degree.  The bounds from your .xml file give 1x0.8 degrees, or 20x16 pixels.

Hi Zhigang,

Please let us know the input file name and parameters for l2gen so that we can generate your input file for Mosaic here.  Both GUI and command line mosaic use the same implementation of the mosaic operator. The only difference could be the way default values are used in the final execution. If this is the case,  George's recommendation, which is to include pixel size variables in your xml, should fix the problem.

Once we generate your input file, we will investigate the discrepancy between the outputs of GUI and command line mosaic tools.

Thanks,
Aynur

Hi Aynur,
The problem has been fixed according to George's recommendation. I thought the pixel resolution was determined automatically according to different sensors(250m,300m,30m etc.) like the reproject operation. Now the mosaic works well.
A file including the input file name and parameters for l2gen was uploaded as a attachment, maybe it was unnecessary now.

Regards,
zhigang
attachment 1

Thanks George, the problem has been solved according to your recommendation.
I had thought the pixel resolution was determined automatically like the reprojection tool, but that was wrong.
I must assign a fixed value to execute mosaic oration, if not, the default pixel resolution 0.05 would be adopted.

Regards,
zhigang