SAR Simulation Difficulties

Hi, I’m working with ALOS2 1.1 data to create a layover and shadow mask. I have calibrated and filtered my data. I’m successful at orthorectifying my data using the Doppler Terrain Correction, but every time I try to do the SAR Simulation or the SAR Simulation Terrain Correction I run into errors. Here’s the most recent one:

A problem occurred during the target product processing
Type: OperatorException
Message: 0

I really don’t know what to do with this. I’ve tried doing it with a subset of the data after Doppler Range TC, but it still runs into errors.

My goal is to get things going with simulation and study the layover/shadow of my dataset. Is it possible that the dataset I’m using doesn’t have enough layover or shadow to work with and this is why I keep running into errors? Rather than focusing on the layover/shadow SAR simulation, should I instead be working on the band maths? If so, are there some specific references I can go through for some direction?

I would appreciate any insight at all into this subject. The manual doesn’t have enough detail and there aren’t any tutorials that I have found that have been helpful.

Thank you!

As far as I understood it, the SAR-Simulation Terrain Correction only works on non-corrected data. No need to apply RD TC before.

Thank you ABraun for your reply. I continued working the problem and was able to solve the processing difficulty. I saved the Target file in a different folder from the Source file. Besides this, I followed all the same steps as the Georectification Tutorial describes.

Now that I have the masks available, however, I can’t seem to analyze the data. Everything is black and the values are all zero. I converted the bands to dB and had the same result. When I move my cursor over the map and view the results with the Pixel Info there isn’t anything to see. I’ve tried opening the bands with the histogram and have the same story. Every pixel has the value of 0.

When I do the Doppler Range Terrain Correction I have valid data, but for my particular task, I need the SAR Simulation tool.

Can you offer some insight? Anything would be greatly appreciated.

Thank you,

if the no data raster is only the result of of the SAR Simulation terrain correction, the processor was not able to match the patterns of your image with the simuated one. Did you try the other SRTM as well (1Sec=30m, 3Sec=90m)?
You can try to modify the parameters of the cross correlation, e.g. increase the number of GCPs.

If you only need the shadow / layover mask you can as well perform the SAR Simulation (without TC) and then apply the Range Doppler Terrain Correction on your simulated products (including the masks)

Thank you for the reply and good advice. I’m following your instructions and hopefully will have full results by tomorrow. Is it necessary that the SRTM be the same from the Simulation to the Range Doppler Terrain Correction? That is, if I set the SAR Simulation with SRTM 1Sec, does the RDTC have to also be 1Sec or can it be 3Sec?

So far I have found that the SAR Simulation layover-shadow mask has values of -8.98 - -9 when I use the logarithmic scale on the histogram. Does this make sense in your experience? This value is for every single pixel in the mask.

Thank you for your help in this trouble-shooting,

No, you can chose which ever you want. The first one determines the accuracy of your shadow/layover masks and the second determines how good your data is geocoded. It makes sense to use the same for both but it is not necessary.
The shadow and layover mask should actually be a binary raster with those area selected which are affected by the corresponding geometric effects. The actual simulated image can have backscatter variations between -35 and 5 db.
I just tested it and for my case (TerraSAR-X) no mask was created.

Great! Thank you. I was able to get a Simulated Intensity HH + a layover_shadow_mask_HH as the output after completing the SAR Simulation and then the Range Doppler Terrain Correction. However when I open the “mask” there aren’t any separations from layover to shadow or shadow + layover. There aren’t any band masks. Is this what you mean when you say that you tested it and no mask was created??

I didn’t receive the mask file at all, so that is already a plus. I am not sure how exactly it is coded but I can remember that someone once mentioned in here. Maybe you find it via the search function.

Thanks Braun for all of your help. I was able to output the layover-shadow masks as Geotiff’s for several datasets. Now I’m working with them in ArcMAP and trying to see what I can do. Thank you.

good job - glad I could help!

Hi ABraun,
I was hoping you might be able to help me more regarding the mask values. I recently worked with some newly acquired imagery of the Alps. The 96% majority of my mask values were 1, which according to the SNAP docs, should line up with layover. There are also some points with values of 2 (0.036%) and 3 (0.006%). And if I zoom in on the values in the histogram, you can see that there are values until 10. In comparison to 1-3, clearly these are minimal, but what ARE these values? What do they represent? Looking at the code it seems that the ONLY values I should receive at any time using this tool should be 0-3. WHY and WHAT do the 4-10 values represent?? It doesn’t make sense to me. Thank you in advance!

I don’t know, to be honest.

According to the module info it was developed by @junlu and @lveci. Maybe they can clarify

Ok, well thank you for the clarification. I’ll see if I can grab their attention. @junlu @lveci

Dear Mr. Veci and Junlu,

I would very much appreciate it if you could help explain why when I do the Shadow Layover Mask the results include values different than 0-3. I have used the mask on several datasets from different global locations. I always get the 0-3, but I also receive results that are negative numbers and positive >3. Can you please explain what these values represent and why they are occurring?

Thank you very much,