area measurement problems

[J.C. Steele]

Hello all,

  I am helping a friend of mine on the technical aspects of measuring the
area for a project she is working on.  I am having some problems with the
measurement though....what we have is a collection of photos (about 900) of
mussel's and need the measure the overall area of the shell, and then the
area of the damaged portion of the shell (the white portion in the picture
referenced below).  I have been using the tutorial at
http://rsb.info.nih.gov/ij/docs/pdfs/examples.pdf to help me measure the
area, however I have ran into some problems I need some assistance with.
First off, most (if not all)  of the specimens collected have additional
"material" attached the shells that i believe is being included in the
measurements...mine example is at
https://ww2.coastal.edu/jcsteele/imagej/10A.JPG which on the bottom side of
the shell you can see some grasses, etc.  Is there a way to remove this from
the image, or exclude it from the measurement?  As you can see from
https://ww2.coastal.edu/jcsteele/imagej/make_binary.jpg and
https://ww2.coastal.edu/jcsteele/imagej/show_outline.jpg when I issue the
"make binary" command from the tutorial, alot of a of the excess material is
being thrown into the mix when calculating the area (I believe).  Any ideas
on where i should start?

Just for the sake of telling everyone, I do not have much experience with
imagej, however my friend was using it and using the selection tool to
select each individual pixel around the shell, then around the damaged
portion of the shell to make a measurement....this is painfully long, and
with as many photos as she has would take along time...this is when I found
the tutorial (before she was not setting a measurement scale or anything
which I believe would cause severe problems and told her...this is when she
asked me for assistance.)

Any help or comments at all would be appreciated and welcomed, Thanks.


Josh

[Rotella, Anthony M. (GRC)[]]

JC,

        One idea is to mask this off and use the image calculator to rid
the image of it. For instance, in the original image, take the freehand
selection tool, zoom in as much as necessary, and select the grass(only
the interface needs to be accurate). Then go to
Edit-->Selection-->Create Mask. Now do Process-->Image Calculator, and
add the original and Mask. This should White out the grass. Now
threshold, and it should blend in with the paper behind it. If you don't
have to do so many of these, it's no biggie. There may be better ways of
doing it, but this should work.

Tony

>-----Original Message-----
>From: ImageJ Interest Group [mailto:[hidden email]] On
>Behalf Of J.C. Steele
>Sent: Thursday, July 12, 2007 3:02 PM
>To: [hidden email]
>Subject: area measurement problems
>
>Hello all,
>
>  I am helping a friend of mine on the technical aspects of
>measuring the area for a project she is working on.  I am
>having some problems with the measurement though....what we
>have is a collection of photos (about 900) of mussel's and
>need the measure the overall area of the shell, and then the
>area of the damaged portion of the shell (the white portion in
>the picture referenced below).  I have been using the tutorial
>at http://rsb.info.nih.gov/ij/docs/pdfs/examples.pdf to help
>me measure the area, however I have ran into some problems I
>need some assistance with.
>First off, most (if not all)  of the specimens collected have
>additional "material" attached the shells that i believe is
>being included in the measurements...mine example is at
>https://ww2.coastal.edu/jcsteele/imagej/10A.JPG which on the
>bottom side of the shell you can see some grasses, etc.  Is
>there a way to remove this from the image, or exclude it from
>the measurement?  As you can see from
>https://ww2.coastal.edu/jcsteele/imagej/make_binary.jpg and
>https://ww2.coastal.edu/jcsteele/imagej/show_outline.jpg when
>I issue the "make binary" command from the tutorial, alot of a
>of the excess material is being thrown into the mix when
>calculating the area (I believe).  Any ideas on where i should start?
>
>Just for the sake of telling everyone, I do not have much
>experience with imagej, however my friend was using it and
>using the selection tool to select each individual pixel
>around the shell, then around the damaged portion of the shell
>to make a measurement....this is painfully long, and with as
>many photos as she has would take along time...this is when I
>found the tutorial (before she was not setting a measurement
>scale or anything which I believe would cause severe problems
>and told her...this is when she asked me for assistance.)
>
>Any help or comments at all would be appreciated and welcomed, Thanks.
>
>
>Josh
>