search

e-sensing / sits

Satellite image time series in R
https://e-sensing.github.io/sitsbook/
GNU General Public License v2.0
456 stars 76 forks source link

Dealing with invalid areas in SITS #266

Closed gilbertocamara closed 3 years ago

gilbertocamara commented 3 years ago

Many image tiles cover ocean areas, where their values are set to NA. This is the case for Sentinel-2, Landsat, CBERS and MODIS. SITS needs to include a default land-sea mask to avoid wrong classification values.

rolfsimoes commented 3 years ago

Currently, sits is converting full NA time series to zero values which avoids an error.

This is not an ideal solution, but at least enables users to classify regions next to large water bodies, like oceans.

I think sits could automatically generates a mask indicating with how many valid values (i.e. no clouds, no shadows, no NA) each pixel was classified.