incomplete particle analysis

[uetli]

Hi everybody,

I use ImageJ to analyse the air bubble distribution in volcanic rocks. For that I digitized the bubbles, so that I have the outline of the particle as a black line and black areas which are the bubbles. Now I wanna know the area, major and minor axis of the bubbles. So I make the picture binary and chose "analyze particles".  Unfortunately not all bubbles get analyzed... Some do, some dont, and I dont understand what would be the reason for either. I tried to redraw some of them with illustrator, sometimes it solves the problem, mostly it doesnt. I added the 2 pictures ( one of the particle with the bubbles and one of them after processing with iJ) to illustrate the problem. Does anybody know why some bubbles don't get included.

Also I noticed that if I remove the outline of the particle, hence only have the bubbles to analyse, they ALL get analysed as wished. But as I also need the data of the whole particle, it would be way handier if I could analyse the bubbles as being holes in the big particle...

I'm happy for any suggestions or ideas

uetli

[Michael Schmid]

Hi Uetli,

you shoud to separate the outline and the image of the bubbles (separate images, or separate slices of a stack).  Otherwise, the outline will be always considered part of the bubbles where it touches.

If you want to split particles that touch (bubbles that have merged), use Process>Binary>Watershed. As an alternative, to split only those almost separate, erode the image (best with Process>Filters>Minimum) by a few pixels and add the pixels eroded to the semi-axes of the ellipses.

Michael
________________________________________________________________
On May 7, 2012, at 00:22, uetli wrote:

> Hi everybody,
>
> I use ImageJ to analyse the air bubble distribution in volcanic rocks. For
> that I digitized the bubbles, so that I have the outline of the particle as
> a black line and black areas which are the bubbles. Now I wanna know the
> area, major and minor axis of the bubbles. So I make the picture binary and
> chose "analyze particles".  Unfortunately not all bubbles get analyzed...
> Some do, some dont, and I dont understand what would be the reason for
> either. I tried to redraw some of them with illustrator, sometimes it solves
> the problem, mostly it doesnt. I added the 2 pictures ( one of the particle
> with the bubbles and one of them after processing with iJ) to illustrate the
> problem. Does anybody know why some bubbles don't get included.
>
> Also I noticed that if I remove the outline of the particle, hence only have
> the bubbles to analyse, they ALL get analysed as wished. But as I also need
> the data of the whole particle, it would be way handier if I could analyse
> the bubbles as being holes in the big particle...
>
> I'm happy for any suggestions or ideas
>
> uetli  http://imagej.1557.n6.nabble.com/file/n4956499/J_1_23.jpg 
> http://imagej.1557.n6.nabble.com/file/n4956499/Drawing_of_J_1_23.jpg 
>
> --
> View this message in context: http://imagej.1557.n6.nabble.com/incomplete-particle-analysis-tp4956499.html
> Sent from the ImageJ mailing list archive at Nabble.com.

[Herbie-6]

Good morning uetli,

please realize that the surrounding line in fact creates an enormous
particle. Breaking up this line at strategic locations may give you
an impression of what I mean.

HTH

Herbie

>Hi everybody,
>
>I use ImageJ to analyse the air bubble distribution in volcanic rocks. For
>that I digitized the bubbles, so that I have the outline of the particle as
>a black line and black areas which are the bubbles. Now I wanna know the
>area, major and minor axis of the bubbles. So I make the picture binary and
>chose "analyze particles".  Unfortunately not all bubbles get analyzed...
>Some do, some dont, and I dont understand what would be the reason for
>either. I tried to redraw some of them with illustrator, sometimes it solves
>the problem, mostly it doesnt. I added the 2 pictures ( one of the particle
>with the bubbles and one of them after processing with iJ) to illustrate the
>problem. Does anybody know why some bubbles don't get included.
>
>Also I noticed that if I remove the outline of the particle, hence only have
>the bubbles to analyse, they ALL get analysed as wished. But as I also need
>the data of the whole particle, it would be way handier if I could analyse
>the bubbles as being holes in the big particle...
>
>I'm happy for any suggestions or ideas
>
>uetli  http://imagej.1557.n6.nabble.com/file/n4956499/J_1_23.jpg
>http://imagej.1557.n6.nabble.com/file/n4956499/Drawing_of_J_1_23.jpg
>
>--
>View this message in context:
>http://imagej.1557.n6.nabble.com/incomplete-particle-analysis-tp4956499.html
>Sent from the ImageJ mailing list archive at Nabble.com.

[uetli]

Michael Schmid-3 wrote

you shoud to separate the outline and the image of the bubbles (separate images, or separate slices of a stack).  Otherwise, the outline will be always considered part of the bubbles where it touches.

So now I use seperate images for the outline and the bubbles which works fine considering that now all bubbles get counted. The problem is when I compare the areas ( i.e the sum of the bubble areas to the area of the whole particle) the area of the particle is smaller ! The scale I set and the picture quality should be the same in both as far as I know ... so how can this happen ? :-/

[Anderson, Charles (DNR)]

Run analyze particles with the include holes feature to get the total area. Without it, you are getting the area of the outline itself, or the total-bubbles, depending on what image you analyze.

-----Original Message-----
From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of uetli
Sent: Monday, May 07, 2012 6:55 AM
To: [hidden email]
Subject: Re: incomplete particle analysis

Michael Schmid-3 wrote
>
>
> you shoud to separate the outline and the image of the bubbles
> (separate images, or separate slices of a stack).  Otherwise, the
> outline will be always considered part of the bubbles where it touches.
>

So now I use seperate images for the outline and the bubbles which works fine considering that now all bubbles get counted. The problem is when I compare the areas ( i.e the sum of the bubble areas to the area of the whole
particle) the area of the particle is smaller ! The scale I set and the picture quality should be the same in both as far as I know ... so how can this happen ? :-/


--
View this message in context: http://imagej.1557.n6.nabble.com/incomplete-particle-analysis-tp4956499p4957608.html
Sent from the ImageJ mailing list archive at Nabble.com.