Surface cracks analysis

[Dav]

Hello,

I am kinda new in the field of image processing and analysis and was wondering if anyone could help me out.
What I am trying to do is calculate the average area fraction of cracks in this image from cropped samples shown further below.



 

However, as you can see in the images, there are some dark regions on the surface which are not cracks and after applying a threshold level to the image, I have to manually remove these regions each time.
Also if I decrease the threshold level a bit lower, most of the unwanted dark regions will disappear but some cracks will disappear as well.
It is quite time consuming for me as I have over 200 images to analyse from different specimens.

Is there a way to reduce the amount of cleaning I have to do on each image or are there other processing techniques that will make the cracks stand out more clearly?

Thanks! :)

[fmonson]

This reply mayl not provide help for specimens already imaged, unless the specimens are available to be reimaged, but it may cut down on work in future.

http://onlinelibrary.wiley.com/doi/10.1002/sca.4950190605/pdf
http://www.nemati.com/pdf/ijrmms.pdf
http://www.nemati.com/pdf/rilem.pdf
http://faculty.washington.edu/nemati/chapter1.pdf
http://faculty.washington.edu/nemati/chapter2.pdf

http://etd.fcla.edu/UF/UFE0008824/shah_v.pdf

Wood's metal (w/Pb and Cd) or Field's Metal (no Pb or Cd) are Bi alloys (low melting points) that have been used to define micro-fractures in concrete.

With SEM/EDS, one can map elements not in the sample and analyze with relative ease - see URLs above.

Hope this helps,

Fred Monson

[kirill8910]

Hello, Dav!

I spent a lot of time trying to filter out surface cracks from wood surface and still doesn't have any good solution. Are the noise particles that you have to remove less in size then the real cracks? Because one of easy solutions could be use particle remover - everything that less than certain size will be removed (You can record a macro for that). Also I see that cracks are sharper in shape than unwanted dark objects. So maybe think in this direction, for example use advanced threshold plugins, like "trainable segmentation". You can download it.

Kirill

07.11.2012, 18:56, "Dav" <[hidden email]>:

> Hello,
>
> I am kinda new in the field of image processing and analysis and was
> wondering if anyone could help me out.
> What I am trying to do is calculate the average area fraction of cracks in
> this image from cropped samples shown further below.
>
> <http://imagej.1557.n6.nabble.com/file/n5000717/kfKN6.jpg>
>
> <http://imagej.1557.n6.nabble.com/file/n5000717/aIBwb.png>
> <http://imagej.1557.n6.nabble.com/file/n5000717/hnTNZ.png>
>
> However, as you can see in the images, there are some dark regions on the
> surface which are not cracks and after applying a threshold level to the
> image, I have to manually remove these regions each time.
> Also if I decrease the threshold level a bit lower, most of the unwanted
> dark regions will disappear but some cracks will disappear as well.
> It is quite time consuming for me as I have over 200 images to analyse from
> different specimens.
>
> Is there a way to reduce the amount of cleaning I have to do on each image
> or are there other processing techniques that will make the cracks stand out
> more clearly?
>
> Thanks! :)
>
> --
> View this message in context: http://imagej.1557.n6.nabble.com/Surface-cracks-analysis-tp5000717.html
> Sent from the ImageJ mailing list archive at Nabble.com.
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Dav]

Thank you fmonson and Kirill :) Will definitely look into it