[vegetation monitoring] really bright parcels in VV polarization

[ravisik]

Hey,

(I’m not sure that my post is directly related to this forum, I can delete my post if necessary).

I’m currently working on S1 SAR images for vegetation analysis (maize parcels). I observed that few parcels are much brighter than the others regarding their VV polarization. The thing is that looking at a S2 image, I can’t see any difference that could explained why this is occurring. Note that this phenomenon seems to occur with bare ground only so I am suspecting an interaction between S1 and the ground. Also, it occurs at different acquisition dates (not only once), consequently it can’t be a “coincidence”.

I put an example, where the parcel at the left is clearly brighter than the others. This is a composite image (Red is linear VH and Green is linear VV). Values are close to 30, instead of ~0.3 for the normal parcels. I can give you more information/images if necessary.

Thanks a lot,
Florian

[falahfakhri]

I think this what you’re looking for,

Creating Radar Vegetation Index

It will not completely replace NDVI because it is based on surface characteristics (mainly the ratio between cross-polarization and co-polarization) instead of the chlorophyll concentration. But similar to the NDVI higher values indicate higher vegetation presence. It is quite a relative measure with no distinct minimum or maximum.

And this

Creating Radar Vegetation Index

I’d say there is no right or wrong. If you technically derived the index according to the published equation (make sure to set paranetheses around the numerator and denominator), the index is correct. While for some vegetation it might match the variation of the NDVI, it coudl be different for others. It is important to understand that all information derived from SAR backscatter originates from roughness, size and orientation of the objects or surfaces within a pixel.

[ravisik]

Thanks!

Actually I’m looking for a physical explanation (or more an agronomic explaination) of this phenomenon.

The simplest explanation could be the form of plowing, but as I’m not a SAR specialist maybe other parameters could interact.

[mengdahl]

It could be the orientation of the plowing causing high backscatter at some observation angles.

[ravisik]

Thanks! It seems the most probable explanation indeed, I wanted to be sure (maybe some people working on that field are used to deal with this situation).

[mengdahl]

Here’s a good related reference:

https://ieeexplore.ieee.org/abstract/document/1610816

[ravisik]

Really interesting, thanks a lot!

Another study few years later regarding this topic : https://ieeexplore.ieee.org/document/6351968