Radiometric & Geometric Correction Workflow

I understand the theoritical difference between beta0 and sigma0, if I want to do intensity correlation analyisis between 2 images, or in this case 3 images. I was usggested to use sigma0, however the area I am working is very hilly and has some mountains. The values of sigma0 and beta0 are ofcouse different from pxel to pixel, but only beta0 allows me to do terrain flatering.

What do you suggest I can do for my application in this case?
Thank you

I then would try to correct for Gamma0 which normalizes topographic effects.
So first convert to Beta0 and then run the radiometric terrain flattening which produces ‘flattened Gamma0’.

1 Like

I want to work with Sentinel 1 … should I use these steps?

Excuse me, my information is low. I can do the preprocessing process with the header. Help me more

how about starting with some tutorials first?


4 posts were merged into an existing topic: How to obtain backscatter from sentinel 1 A product?

The s1tbx can be used to process Sentinel-1A/B data. Does the s1tbx is open source? We want to apply some new algorithms to extract Sentinel-1A/B SAR information, thus the input of Sentinel-1A/B is necessary. Therefore, can you supply us the Sentinel-1A/B input source codes or other approach to obtain the Sentinel1A/B import codes? That is to say the sentinel-1A/B can be read. Thank you very much.

1 Like

Otherwise there is the option of snappy with/or:


I want to know whether I could do thermal noise removal after applying orbit file and calibration to Sigma0 or not?

Thanks for your reply in advance.

I haven’t found any source, but at least before calibration seems reasonable. Orbit files have a geometric effect in the first place.

1 Like

Hi @lveci @ABraun,
I’m looking for help on terrain flattening for SLC.
Iv been working on change detection using (C2) covariance matrices from stacks of SLC complex images. I am experiencing some false positives over slope. Pipeline is: Orbit correction --> Radiocalibration (to complex) -->Deburst–>Multilook–>Terrain Correction.

As discussed, the GRD pipelines can normalize to gamma0 (e.g. radiocal to beta0 --> TF), but this option isn’t obviously available for complex. In fact, any normalisation under radical is not available if complex output is selected. Is there another tool like TF that i’m missing?
Much appreciated,

You are right, Terrain Flattening currently only accepts Beta0 as an input.
Maybe you can apply Range Doppler Terrain Correction on your c2 matrix with the ‘apply radiometric normalization’ checkbox selected.

I haven’t tried it in this context but it has similar effects as Terrain Flattening.
Maybe you test with and without and use a scatterplot to see if it has an effect.

Great, i also missed the option to output in complex which doesn’t cancel out the normalisation . Ill have a look and get back to you

Right, the complex output at the TC step is also quite new. I would be interested in your findings.

Per chance do you know roughly when it was added?

Revisited this today and found that checking the radiometric normalisation (in the TC) section flagged a conflict warning that it could not be used with polarimetric matrix product :frowning:

@SAR2016 did you code ortho algorithm by yourself or you get it from some online database? Is there any database from where we can python code for SAR processing algorithms?

Hello sir,
Does it mean, if we are performing terrain flattening then there is no need to perform Range doppler TC??

They do different things:
Terrain Flattening is a correction of radiometric distortions distcaused by topography. It reduces the effect of topography on backscatter intensity.
RD Terrain correction is a correction of geometric distortions caused by topography, such as foreshortening and shadow. It corrects the location of each pixel based on a digital elevation model.