[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


Mark Chan (chanm@cadvision.com) writes:

> An image consisting of two rock types. Rock type 1 is fragmented, dispersed
> irregularly and embedded within rock type 2. Type 1 is generally dark brown
> but not of unique color, type 2 color varies between black or dark brown in
> color.
> Want:
> The fraction of type 1 as a percentage of total image.
> Questions:
> 1) Any suggestion on the easiest way to do such computation?
> 2) Will some sort of image pre-processing help get the job done easier?
> 3) Any routine/program in existence I can utilize?

I've built a similar sort of "classifier" for a client.
We have had pretty good luck taking the 24-bit image
and converting it to hue, value, and saturation images.
Then I wrote a little interactive gizmo (based on my
XStretch program) that allows us to independently
select pixels based on hue and value. Pixels that
lie in both camps get selected as the "thing" we
are trying to classify. 

We have found that we can get pretty good segmentation 
with this approach, at least for the data we are interested 
in. Your mileage may vary. 



David Fanning, Ph.D.
Fanning Software Consulting
Phone: 970-221-0438 E-Mail: davidf@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155