Polarimetric Matrix

s1tbx
#1

I am attempting to create H/alpha plane plots for a number of RADARSAT-2 quad pol datasets. My initial workflow is as follows:

read >> radiometric calibration >> polarimetric speckle filter >> polarimetric decomposition >> terrain correction >> write

…however, I have noticed that in the batch processing tool in SNAP there is the option to run a H-a Alpha classification which includes the following step:

read >> calibration >> polarimetric matrices >> polarimetric speckle filter >> polarimetric classification >> write

My first question is, are these 2 models different? I understand one is the decomposition and the other a classification but both are generating an H-a Alpha dataset. My second question is, if they are similar, do I need to include the polarimetric matrices in my initial workflow? My final question, can you provide insight into the matrix generation?

Cheers,

Matt

#2

They are different. The classification is a Wishart classifier using H-a Alpha.
If you don’t use pol matrices, it will automatically convert to T3. You would use pol matrices if you wanted something other than the default T3.

1 Like
#3

Thank you so much for the explanation Iveci, that is exactly what I was looking for.