I have been trying to mask out various classified objects within S2 L2A data. First, I resample the product, then try to use the Land/Sea Mask to apply a mask of multiple data: the scl_nodata, scl_saturated_defective, scl_dark_feature_shadow, scl_cloud_shadow, scl_water, scl_cloud_medium_proba, scl_cloud_high_proba, scl_thin_cirrus and scl_snow_ice. My problem is I am unable to apply a second round of masks after applying the first. I have tried to do this in the graph tool but I receive a bufferdata error. I have read this may be due to an insufficient amount of ram. I have also tried to do this in band math, but I don’t know if you can apply the math to the original band which would allow you to use in the biophysical processor. I am wondering, how does one mask out multiple objects from the scene classification as I have tried using gpt in the command line as well as in the visual graph builder and it does not seem possible.
There is a Mask Manager that allows you to create new masks with a combination of several masks of the product.
2 Likes