I’ve noticed that, with the the release of SNAP 8.0, the new version of the Biophysical Processor differentiates between S2A and S2B as input into the neural network. It is stated that ‘this results in output more reliable and with better accuracy and takes into account the small differences in the reflectance response of the different sensors’.
Do have any idea about the magnitude of accuracy improvement that we look at? Is a comparison between the results of the old and new version available?
I’m trying to determine whether it is worthwhile to reprocess the archive of LAI images that I am currently working with.
I do not have the exact number (I will try to find some number in the next days) however the spectral responses of the S2A and S2B satellites have some differences as you can see in the graph below and the previous model was tuned only on the S2A sensors (the S2B data were not yet available when the previous version of the algorithm has been developed).
