Detect clumps (patches) of connected cells. Each clump gets a unique ID. NA and zero are used as background values (i.e. these values are used to separate clumps). You can use queen's or rook's case, using the directions argument. For larger files that are processed in chunks, the highest clump number is not necessarily equal to the number of clumps (unless you use argument gaps=FALSE).
Usage
# S4 method for RasterLayer
clump(x, filename="", directions=8, gaps=TRUE, ...)
Arguments
x
RasterLayer
filename
Character. Filename for the output RasterLayer (optional)
directions
Integer. Which cells are considered adjacent? Should be 8 (Queen's case) or 4 (Rook's case)
gaps
Logical. If TRUE (the default), there may be 'gaps' in the chunk numbers (e.g. you may have clumps with IDs 1, 2, 3 and 5, but not 4). If it is FALSE, these numbers will be recoded from 1 to n (4 in this example)