Author: | Aaron |
Description | An experimental plugin for performing k-means clustering of RGB image files. This can be useful as a pre-filter for importing attributes maps. |
Menu option | 'Extensions→atCluster→Cluster image' |
Download link | atCluster-bin.zip [140kB] |
Last update | 31 May 2015; v15.5.0.1 |
How to install? | Please refer to the plugin installation guide. |
This is an experimental plugin. The user interface is still very basic. Not recommended for production environments.
Select 'Extensions→atCluster→Cluster image'.
Select an import image.
Enter settings
nClustersMax is the maximum number of clusters to be found. If importing for an attributes map, keep this small (« 100).
nIterationsMax is the maximum number of iteration cycles in the k-means algorithm. 10 is recommended.
atCluster will then cluster the image and export a file called '[orignal file name]_clustered.bmp'. This file should look the same as your input image, but with slightly 'stepped' colours. This output image will only contain a small number of unique colours (as given by nClustersMax), which should be suitable for importing as an attributes map.
Select '
Import→Attributes map→Import from colour map', and follow
this guide. Please note that you will need to manually assign climates/land types to the colours in the image.
2015/05/31 06:07 — Released v15.5.0.1.