Classifies a block of data using multicores. It breaks the data into horizontal blocks and divides them between the available cores.
Reads data using terra, cleans the data for NAs and missing values. The clean data is stored in a data table with the time instances for all pixels of the block. The algorithm then classifies data on an year by year basis. For each year, extracts the sub-blocks for each band.
After all cores process their blocks, it joins the result and then writes it in the classified images for each corresponding year.
.sits_classify_multicores(
tile,
ml_model,
roi,
filter_fn,
impute_fn,
interp_fn,
compose_fn,
memsize,
multicores,
output_dir,
version,
verbose
)a single tile of a data cube.
model trained by sits_train.
region of interest
smoothing filter function to be applied to the data.
impute function to replace NA
function to interpolate points from cube to match samples
function to compose points from cube to match samples
memory available for classification (in GB).
number of cores.
output directory
version of result
print processing information?
List of the classified raster layers.