Effect of Multilooking on Terrain Flattening

I am working on processing SAR scenes for a mountainous area, and the highest resolution DEM I have access to for the area is at 1 arc-second. The processing flow I’d like to employ is:

-Deburst -> Remove Thermal Noise (if applicable) -> Apply Orbit File -> Calibrate -> Speckle Filter -> Terrain Flatten -> Multilook -> Terrain Correct

As the DEM is of coarser resolution than the SAR scene, when I try to terrain flatten SNAP posts a warning that my DEM is of coarser resolution than the SAR scene and I should multi-look the product before Terrain Flattening.

As Small’s paper on flattening gamma focuses on radar geometry interacting with topography to influence both the radar cross section as well as returned brightness, I am curious how multi-looking will affect the accuracy of my Terrain Flattened brightness values. Does anyone have thoughts on the matter?

Thanks in advance,

You can skip the multi looking by changing the re-grid option in the terrain flattening module.

A comparison of the result with and without multi looking would be interesting.

Terrain flattening without multi looking, at least using the SRTM 1 arc-second HGT DEM, results in large blank portions of the image:

This effect carries over into the final product, and I’m unsure why. If I multi-look first, the blank portions are mitigated. I’ll try with an Aster 1 arc-second DEM as well, but am curious if anyone has had similar issues? As multi-looking is recommended through the software prior to Terrain Flattening, is there also documentation on how multi-looking might impact the accuracy of the brightness values on the processed scene?

Looks like you may have had ‘Mask out areas without Elevation’ selected? That’s usually the cause of a black mask like that :slight_smile: