🚀 go-pugleaf

I have an experiment from which I am trying to extract information by way of 
digitized video and I am having trouble finding routines to do what I want 
to do, almost certainly because I don't know where to look and so I was hoping 
somebody may be able to help me. I have an image of the leading edge of a flow
which is made up of a superposition of eddies, i.e. circles, of varying radii.
What I want to extract is a kind of spectrum of eddy sizes so that I can study
how these develop as the flow progresses. Firstly I need to retrieve the 
underlying curves from the digitized data - surely this must be a trivial form
of curve fitting where it is known that the real curve lies within half a pixel
separation of the digitized 'curve' - all I can find are routines for 
interpolation of data from regular grids, such as a regular array of
measurement points, which is not what I want at all. Then I need to process
the smoothed (piecewise) curves to extract the range of scales and get a kind
of probability density function for sizes. I have a few ideas on how to
accomplish this latter part, but any suggestions would be extremely helpful.
Thanks in advance,
                  Ron Cresswell

Dr R W Cresswell
Centre for Dynamical Meteorology
Monash University
Melbourne
Australia
cressrw@vaxc.cc.monash.edu.au