“/by zero error” occurs in Terrain-correction(EPSG:4326)

[choin0207]

Dear all
I am working with S1 GRD images, but I got error message when I set Terrain-correction’s Map Projection in EPSG:4326.

“/by zero error” shows when I do Terrain-correction (EPSG:4326),
I know there is a already a similar discussion on the forum
by-zero-error-and-sarsim-terrain-correction-error-during-graph-building
but the discussion stop in thomakra’s post:
However, it works if I specify the mapProjection string as:
GEOGCS[“WGS84(DD)”,
DATUM[“WGS84”,
SPHEROID[“WGS84”, 6378137.0, 298.257223563]],
PRIMEM[“Greenwich”, 0.0],
UNIT[“degree”, 0.017453292519943295],
AXIS[“Geodetic longitude”, EAST],
AXIS[“Geodetic latitude”, NORTH]]

So …
Can EPSG:4326 work correctly in Terrain-correction ?
or
Is there any difference between WGS(DD) and EPSG:4326

MY IMAGE: Subset SA1_IW_GRDH_1SDV
MY WORKFLOW: Read -> Calibration -> Terrain-Correction -> Write

[ABraun]

WGS(DD) and EPSG:4326 are the same, and WGS84(DD) should actually work with any SAR product. Which steps were applied before?

[choin0207]

which steps?
Regardless of what steps I used before, it always got the “\by zero” error When I change the map projection to “EPSG:4326”,
In this post, I only test “EPSG: 4326” by following the steps below
Calibration -> Terrain-Correction -> Write

[ABraun]

Have you tried a different image resampling method?

[choin0207]

I just tested
NEAREST_NEIGHBOUR
BILINEAR INTERPOLATION
CUBIC_CONVOLUTION
BISINC_5_POINT_INTERPOLATION
BISINC_11_POINT_INTERPOLATION
BISINC_21_POINT_INTERPOLATION
BICUBIC_INTERPOLATION

All failed…

[ABraun]

according to this post (Aggregation and Interpolation of Sentinel products - should I use snappy or GDAL tools?) it is related to the digital elevation model. Which one are you using and where is your study area?

[choin0207]

Dear ABraun
My study area is in Taiwan,

I have tried SRTM 3Sec, SRTM 1Sec HGT and External DEM(20mDEM get from Naional Web) before. All of them can work normally in WGS84(DD).

I compare the GPF_Parameter_Files(XML) that create from SNAP Graph Builder with forum suggestions you provide.
Only “pixelSpacingInMeter” is different

But I wonder why I need to set “pixelSpacingInMeter” to 0.0 ?
Is that correct?

Finally, even if I set pixelSpacingInMeter in the xml file to 0
Still appearing Error:/by zero

[ABraun]

I’m out of ideas
10 meters makes more sense to me actually.

[marpet]

Hello, I’m sorry to hear that you have stumbled about this problem.
Can you reproduce this error and attach the log file here.
I hope that I can see the stack trace, to identify where this error occurs.

You can find the log file by selecting from the menu Help / Show Log Directory. Make sure to reproduce the error first and then attach the messages.log file.

Thanks

[choin0207]

Dear marpet
Thanks for your help!
messages.log (77.1 KB)

[marpet]

Thanks for the log file.

Could you have a look at the output product of the calibration step?
Is it of zero size, maybe? The error happens at a location in the code where I wouldn’t expect it. Maybe something is wrong with the input of the Terrain-Correction.

[choin0207]

This time I do it step by step

1. GRD->[calibration] -> Cal_product
2. Cal_product->[Terrain-Correction]->Cal_TC_product

Cal_product looks fine…
But it still shows error message at second step:Terrain Correction(EPSG:4326)

messages.log (83.6 KB)