I want to use the biophysical processor for my project and since i’m new to these topics i’m trying to check if i understood it well.
I read the ATBD paper about this product and i want to see if i’m right about the algorithm.
So the procedure is:
- first a “virtual” dataset about plants and terrain’s features was created,
- with these dataset a radiation value was obtained applying a Radiative transfer model,
- then with these data FAPAR, LAI etc were calculated still with a physical method either,
- an ANN was trained with input: radiation and sensing geometry, output: biophysical parameters previously calculated,
- finally this trained model is used to take as input “real” radiation from satellite’s data and infer estimated biophysical values.
If this’s is correct, now i’m wondering if ever such results have been tested and validated on ground truth observations.
Thanks for help.