From Open Foris Wiki
oft-clump - Connected Component Labelling
Add spatial coherency to existing classes by combining adjacent similar classified areas
Usage: oft-clump <-i inputfile> <-o outputfile> [-b band] [-um maskile] [-h help]
Options: -b band = use determined band of the image -um maskfile = use maskfile and process only areas having mask value > 0
- Oft-clump is meant for separating uniform regions in a class image.
- You may obtain such a class image by using e.g. oft-cluster.bash, oft-kmeans or oft-nn.
- The program looks for similar and adjacent class values in the input image and gives each area an own id.
- NOTE: For the benefit of users running the script using the older version, where the datafiles are based on the file order instead of options -i and -o, the program can still be used that way.
- An example of the aim of this program is seen in the image below.
- Colors represent class values
- After clumping: pixels with identical class values, but are not spatially connected will have a different id
- Get Example data set
- For this exercise following tools are used: oft-clump
- Open your working directory using
- To run the oft-clump we use the Input: landsat_t1.tif, Output: clump.tif:
oft-clump landsat_t1.tif clump.tif