(especially in areas of low backscatter intensity)

In the researching progress,I met some basic questions as follows:
1.waht does the “Thermal noise removal” for ,in another word,what is the main purpose for this function?
is it necessary for dual-polarization decomposition?
2.what does “Radiometric terrain flattening” for ?
3.what does “Range Doppler Terrain Correction” for?

I am looking forward to your introduction.
My best regards.

Hello,

  1. Thermal noise removal should be applied on SENNTINEL-1 data acquired over areas with very low backscatter intensity. As the noise floor of SENTINEL-1 is -22db (i think), any radar signal below that level cannot be capture by the sensor, hence, its just noise.
    In the image below, the bright stripes across the image results from the thermal noise. Hence, thermal noise removal is important process

  2. In terms of radiometric flattening, it is applied for areas where the terrain is not flat. If you want to perform classification for an area with a rough terrain and terrain correction is not applied, then inaccurate results are produced. This is because the terrain variations affect the brightness of the radar return.

  3. Due to variations of the scene and the tilt of the satellite sensor, this leads to distortions in SAR images. Terrain correction compensate for these distortions. This step in SNAP projects your SAR image in WGS84 coordinate system (or any other coordinae system you want) as well. So you have a georeferenced nd terrain corrected image at the end

1 Like

thank you very much for your explaination. And can i ask one more question?
how to identify the thermal noise?and when i run the “thermal noise removal” using S-1 SLC data by SNAP,the i and q band was lost,how should i do to preserve the polarimetric information when i do the “thermal noise removal”?

What are you actually trying to do?
what is the purpose of processing SENTINEL-1 data?
Are you using intereferometric method? (you mentioned SLC data )

Thermal noise removal is usually used in Ground Range Detected products to eliminate the noise. If you are doing a classification in a noisy image, this might lead to higher classification problem.

If you use SLC data for producing DEM or a displacement map, you just use the phase of the data and not the intensity.

1 Like

i want to do dual pol decomposition using the slc data,so what should i do before generate the matrix, i.e what are the pre-processing steps for SLC data used for dual pol decomposition?

From my understanding, you are using polarimetric data.
I am not familiar with polarimetric processing.
There is an option in SNAP where you can do polarimetric decomposition

thanks again

For what concerns thermal noise removal. I guess the following TN will help you:
https://sentinel.esa.int/documents/247904/2142675/Thermal-Denoising-of-Products-Generated-by-Sentinel-1-IPF

Also , the noise removal on apply to detected data (GRD) or SLC after detection