Box Counting Method

[mark.lake at ucl.ac.uk]

Markus,

In answer to your questions about r.boxcount:

1) It was written to work with binary images, not greyscale.  I may be
wrong about this, but as I understand it the boxcounting method is not
directly applicable to greyscale images, since it is basically a
measure of the extent to which space is filled.  If you have greyscale
data then presumably you are more interested in *how* space is filled.
To measure that I think you need the kind of measures described in,
for example: T. Peli, 1990, Multiscale Fractal Theory and Object
Characterization, Journal of the Optical Society of America A, vol. 7,
pp.1101--1112.

2) I've used a shell script that repeatedly calls r.boxcount to
measure variability in the fractal dimension of binary maps via a
`moving window'.  The result is a raster map in which each cell is
coded with the fractal dimension of the window centred on that cell.
This is compute-intensive, but not impossibly so.  I'll tidy the
script up and put it on my site.

Sorry to be so slow in replying - I've been away.

Mark

-------------
Dr. Mark Lake

Institute of Archaeology              Tel: +44 171 391 1535
University College London
31-34 Gordon Square
London, WC1H 0PY	              Fax: +44 171 383 2572

http://www.ucl.ac.uk/archaeology/research/profiles/lake.htm