Darrell -
I am certainly interested in directional stats - plenty of geological
information is given as azimuths (eg strike's of surfaces). Strictly
speaking these represent the intersection of (3D) planes (with the earth's
surface), so there is a stereological problem to overcome too, but getting
the 2D directional stuff sorted out would be a very useful first step.
I have a student working on automatically detecting linear features in
natural images - eg Remote sensing, photographs, photomicrographs - which
would form the input data for this. (He is using image enhancement
followed by an implementation of the hough transform - in AVS at present
but when the method is sorted out he will encapsulate it in some
stand-alone code - has anyone else done this?.) To get the proper "dynamic
range" I think that this would have to be done in the vector domain (raster
cells can only have a choice of 8 directions of neighbours) and I am
expecting to be dealing with vectors defined simply by the coords of the
two end points.
BTW you might like to look at N I Fisher, Lewis & Embleton "Statistical
analysis of spherical data" Cambridge UP, 1987 for the 3D versions of the
stats (I think this also touches on 2D) and there is a nice paper by Fisher
and others "Spatial analysis of 2D orientation data" in Mathematical
Geology, v17, pp177-194 which uses PCA and glyphs to build a system for
interactive analysis (you probably already know about these).
Congratulations on the PhD.
Simon Cox
All the talk of circles reminded me put out a feeler on the following:
does anyone deal with data *on* circles? In other words, does anyone
have basic data that are in the form of directions? (An example of
this type of data may include directions that animals leave a
nest/hole in search of food/mates/etc.) Just for fun-zees, I'm
contemplating writing a program to statistically test for randomness
of directions (testing the hypothesis that the n points are randomly
distributed about the circumference of the circle). I have all
the internal nasties already written (e.g., Kuiper's V and Watson's
U^2 stats) so writing a GRASS wrapper would be pretty easy.Let me know.
--Darrell
James Darrell McCauley, PhD http://soils.ecn.purdue.edu/~mccauley/
Dept of Agricultural Engineering mccauley@ecn.purdue.edu
Purdue University tel: 317.494.9772 fax: 317.496.1115
___________________________________________________________
Dr Simon Cox __ \
CSIRO Exploration & Mining ,~' L_|\ Australian
39 Fairway, PO Box 437, ;-' \ Geodynamics
Nedlands, WA 6009 Australia ( \ Cooperative
Phone +61 9 389 8421 + ___ / Research
Fax +61 9 389 1906 L~~' "\__/ Centre
s.cox@dem.csiro.au 7
___________________________________________________________