Open fronzag opened 3 years ago
Principally yes. Making SpecHomo compatible to a new sensor requires the spectral response functions of the sensor. They have to be added to the pyrsr package which is used by SpecHomo. This allows to train material-specific regressors as described here which can then be used to estimate the target sensor spectra. Based on these spectra, you could compare the purely sensor-induced spectral differences between Landsat-8 and CBERS 4.
Thank you Daniel!
I'll talk with my project leader and if this is a task, I'll talk with you again, ok?
Thanks! José Guilherme Fronza
Em qui., 1 de abr. de 2021 às 10:36, Daniel Scheffler < @.***> escreveu:
Principally yes. Making SpecHomo compatible to a new sensor requires the spectral response functions of the sensor. They have to be added to the pyrsr https://git.gfz-potsdam.de/geomultisens/pyrsr package which is used by SpecHomo. This allows to train material-specific regressors as described here https://geomultisens.git-pages.gfz-potsdam.de/spechomo/doc/usage/classifier_creation.html which can then be used to estimate the target sensor spectra. Based on these spectra, you could compare the purely sensor-induced spectral differences between Landsat-8 and CBERS 4.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/GFZ/spechomo/issues/1#issuecomment-811913452, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB4L7WJM27EYVYSO7BL25QDTGRZG7ANCNFSM42FBEHMA .
Is it possible to add CBERS 4 and CBERS 4A WFI sensor bands to this package? I want to compare CBERS 4 / 4A WFI band results (NIR) to Landsat-8 OLI (NIR).