An implementation of a fast clumping algorithm.

netzel netzel Minor changes in makefile(s) 1da2363 @ 2024-05-16 14:28:15
Makefile Minor changes in makefile(s) 2024-05-16 14:28:15
Makefile.win Minor changes in makefile(s) 2024-05-16 14:28:15
README.md README changes 2024-02-24 19:42:51
clump.c Initial commit 2024-02-21 22:59:53
data.c Initial commit 2024-02-21 22:59:53
data.h Initial commit 2024-02-21 22:59:53
gpl-3.0.txt Initial commit 2024-02-21 22:59:53
main.c Initial commit 2024-02-21 22:59:53
main.h Initial commit 2024-02-21 22:59:53
README.md

Multi-categorical Connected Component Labelling (CCL) algorithm identifies all 4- or 8- connected regions of cells sharing the same categorical values and assigns each of them a unique label.

The plClump4p, a multi-categorical CCL algorithm, is an efficient variant of a two-scan algorithm modified for use in multi-categorical rasters. It applies a divide-and-conquer technique coupled with parallel processing to a standard two-scan algorithm. Using these techniques causes that algorithm's performance is two-three orders of magnitude better than standard CCL algorithms. The specific speed up depends on the size of the data and its complexity and is greatest for very large and complex rasters.

Windows and linux binaries are included in plTools package.

The program is available under GPL v3 license.

The implemented algotithm is described in the following paper: Pawel Netzel, Tomasz F. Stepinski, 2013: Connected components labeling for giga-cell multi-categorical rasters, Computers & Geosciences, Volume 59, pp 24-30, DOI: 10.1016/j.cageo.2013.05.014.