I try to label a huge image (69940, 70935) using ndimage.label. Label thus try to allocate a int64 ndarray (37 GiB) which is too big for my computer.
import rasterio as rio
import ndimage as ndi
with rio.open(src) as f:
label = ndi.label(f.read(1) == feature, structure = struct)[0]
Unable to allocate 37.0 GiB for an array with shape (69940, 70936) and data type uint64
Is there a way to perform this analysis using split chunks ?
