Re: Image analysis - separating touching particles


"rarewolf" <rarewolf@xxxxxxxxx> wrote in message
news:1129483937.072615.189020@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>I am using a particular software for determining particle (and included
> grains) identification. For separating touching particles the software
> offers parameterization of 3 criteria: 'F', for control with regard to
> features across the surface (e.g., cracks); 'T', for controlling
> particles touching at a point; and 'E', for some control of particles
> touching along longer joins. What makes the parameters a bit difficult
> to experiment with is the number of permutations ... That is, the are 3
> parameters for 'F' and 3 for 'E', for 7 total.
>
> Does anyone recognize these parameters, such that they may provide me
> with a reference for understanding parameter values and how they may
> affect one another?
>
> TIA ... michael shaffer :o)
> Memorial University
> St. John's Newfoundland
>

The software sounds a bit complex. In the past there was the
erosion-dilation effect that was used for the "bunch of grapes" effect.
This is were one pixel would be eroded from each object until there were
more objects. This effect was controlled by the operator. If your image
had a series of lightly touching objects this worked Ok, not well.

What may be going on is orientation image conrol, objects can be excluded by
size and orientation. For instance if the crack is large and up and down it
is excluded. The other thing would be recuring forms such as the bunch of
grapes effect but over 10-15 pixels. For this they could be using an
oriented dip to oriented dip approach. Lets say you have a series of large
features oriented in X in a straight line. The program follows the line and
notes when it dips, then it follows the dip until until it starts to go up.
That's the start point. Now it goes down the object until it finds an edge,
a transition point between gray values, this is the end line. If there are
bumps moving up then a point is put on the upper most side. If this point
is within the specified angle then upper and lower dots are connected and a
straight line is draw.

Now all of the above is speculation. Another point is that automatic system
either work real well or you have to putz around with them for ever. You
may be in the for ever group and may find it faster to use the
semi-automatic function that I bet the program has.

These programs aren't all that smart, but they are fast. If that works for
ya, thats great.

Kevin Cunningham
SMS


.

Relevant Pages

  • Image analysis - separating touching particles
    ... For separating touching particles the software ... offers parameterization of 3 criteria: 'F', for control with regard to ...
    (sci.techniques.microscopy)
  • Re: Theory of beliefs (model of browhaha)
    ... plasma, dense, though I do imagine particles as spinning in a plane). ... they define an orientation and can change it more or less ... Spinmasters are ideologic particles, hence their ...
    (sci.econ)
  • Re: Theory of beliefs (model of browhaha)
    ... plasma, dense, though I do imagine particles as spinning in a plane). ... they define an orientation and can change it more or less ... Spinmasters are ideologic particles, hence their ...
    (sci.physics)