Is there a way to measure depth?

[jkaufman]

Hi everyone,

I know that it is very helpful to have an example image to understand what I am asking, however I cannot send out pictures of my materials.  I am working on the same problem as in my previous post, trying to find the area percentage of adhesive left on an aluminum panel.  I start running into thresholding problems when my adhesive is either clear or just a similar color to the aluminum.  
So I am asking if it is possible to measure depth, or just find the area of the part that is furthest away (for me it is the aluminum that is below the adhesive).  The examples I can show you, are sample images from Image J. Take the image boats for example, is there a way to tell which boat is in front, and outline the area of the remaining boats?

Thank you for any help that you can give me even though I am unable to provide you with an actual sample picture.

[Nathaniel Ryckman]

Probably not. Depth perception is a complicated perception created from many measurable quantities and logical inferences.

I see a plugin called IsoPhotContour2, but, since the description isn't very thorough, I'm not sure if it does what you want or not:

http://www.dentistry.bham.ac.uk/landinig/software/software.html

Here is a paper on automated depth perception. Maybe you can pick up some ideas/key words to Google there:

http://robotics.stanford.edu/~ang/papers/aaai08-Make3dDepthPerceptionSingleImage.pdf

I was trying to help someone else with a similar topic. The colors of the object that they were trying to select was too close to the colors of the light shining on it. I had an idea, but I haven't gotten to test it out very much, and it didn't seem to work very well.

How big are the gaps that you are talking about? Close, dilate and fill might be able to help you if the gaps are extremely small.

jkaufman wrote

Hi everyone,

I know that it is very helpful to have an example image to understand what I am asking, however I cannot send out pictures of my materials.  I am working on the same problem as in my previous post, trying to find the area percentage of adhesive left on an aluminum panel.  I start running into thresholding problems when my adhesive is either clear or just a similar color to the aluminum.  
So I am asking if it is possible to measure depth, or just find the area of the part that is furthest away (for me it is the aluminum that is below the adhesive).  The examples I can show you, are sample images from Image J. Take the image boats for example, is there a way to tell which boat is in front, and outline the area of the remaining boats?

Thank you for any help that you can give me even though I am unable to provide you with an actual sample picture.

[jkaufman]

I am following ASTM D1002 standard for my lap shear test, so the approximate area of the joint is 1 in x .5 in.  There are two types of failure of the joint, adhesion failure where the adhesive splits off from the aluminum, and there is a distinct edge between the aluminum and the adhesive.  The problem that I have from these failures is when the adhesive is clear or close in color to the aluminum, and therefore I cannot use thresholding to separate the two parts. Cohesion failure is much more difficult to analyze, as this is a failure of the adhesive, and often there are only small and scattered parts of aluminum visible.

I may be able to solve my problem with adhesion failure by coloring the adhesive with a black sharpie, however cohesion failure is too complex to simply color over.

[Nathaniel Ryckman]

You could give Process->Find edges a try

It's really hard to determine the nature of the problem without seeing an example. Basically, you have to figure out how you are able to see the differences before you can give the computer explicit commands to do the same thing.

Unfortunately, this often has the potential to turn seemingly simple tasks into very complicated ones.

jkaufman wrote

I am following ASTM D1002 standard for my lap shear test, so the approximate area of the joint is 1 in x .5 in.  There are two types of failure of the joint, adhesion failure where the adhesive splits off from the aluminum, and there is a distinct edge between the aluminum and the adhesive.  The problem that I have from these failures is when the adhesive is clear or close in color to the aluminum, and therefore I cannot use thresholding to separate the two parts. Cohesion failure is much more difficult to analyze, as this is a failure of the adhesive, and often there are only small and scattered parts of aluminum visible.

I may be able to solve my problem with adhesion failure by coloring the adhesive with a black sharpie, however cohesion failure is too complex to simply color over.