DEM-Assisted artifacts from gpt

Hi guys,

I would like to inform you about two strange artifacts are arising when using the DEM-assisted operator. I run the operator using gpt with these options:

-PmaskOutAreaWithoutElevation=‘false’ -PtileExtensionPercent=100 -PdemResamplingMethod=‘BICUBIC_INTERPOLATION’ -PresamplingType=‘BICUBIC_INTERPOLATION’

the images are three GRD S1 assembled from track 117 acquired in May 2015. I applied precise orbit, border removal (custom not the one implemented in snap), thermal noise correction and calibration then after the assembly i try to corregister. The master is the largest image.

  1. Even I set “not to mask Area Without Elevation” the sea is masked out.
  2. I got these strange artifacts at the borders of the slave images even though the images are well aligned (before the coregistration the borders are set everywhere to 0).

Moreover, if I use an external dem the process crash with this error:

Error: DEM-Assisted-Coregistration: computeSlavePixPos: java.lang.StackOverflowError

1 Like

Dear @lveci did you have time to check these problems out?

Thanks

C

Dear @lveci did you have time to check these problems out?
they are still persisting in the newer versions

Thanks

C

Could you tell us the data set you’ve used? We will try to repeat the problem and look into it.
Thanks

S1A_IW_GRDH_1SDV_20150517T170617_20150517T170642_005964_007AF1_AEC9.zip

But it happens systematically in almost all the images with different track I pre-processed. (Before the co-registration no artifacts are present)

If you like I can send you the image in dropbox.

Thanks!