Curved Particle Width

[achille]

I have a grayscale image which after thresholding it I have an image like the attached one. As you can see there are particles which are like curved lines.

As you can see the scheme and the orientation of each particle is random.

I want to calculate the width of each particle. Can you please help me?

Thank you.

Achille. 

[Rialter]

Hi;

If you already have a binary image you might find the Particla Analysis function useful. This function does many different analysis on the particules identified in the image, so the average horizontal segment lenght and average vertical segment lenght can be useful for your applicaton. You can use some morphology fuctions like a close functionto fill the holes on the particles if you need so.

I hope the information is useful.

Good luck.

Message Edited by Rialter on 04-30-2009 05:14 PM

[achille]

Dear Francisco,

Thank you for your answer. By calculating the average horizontal and vertical segment length how can I calculate the width of the particle?

As you have seen in the picture the orientation of the particle will be random and not in vertical or horizontal position.

So why the average horizontal and vertical segment length will give me the width of the particle? 

Thank you.

Achille 

[BCho]

Hey achille-

Which part of the particle are you interested in finding the width? On your example you have two particles with regions that have different 'widths.' Is there a specific location on the particle the you wish to measure? Are these particles going to be random places and orientations in subsequent images? If not you could try using a Find Edge step and simply measure the distance between the two edges of your image. Otherwise, you will need to be able to track the orientation of the particle and then find the location where you wish to find the width.

Perhaps a little more clarification on what part of the object you wish to measure would be helpful.

Hope this helps.

-Ben

[achille]

Hi Ben,

Thank you for your message. The particles will be in the picture in random places and orientations.

I would like to be able to measure 2 things:

1. To measure the mean width of the whole particle

2. To measure the width at the beginning of each particle. For example to measure the width at the right beginning of the upper particle and to measure also the width at the left beginning of the lower particle.

Can you please be a little analytic of how can I achieve it?

Than you.

Achille 

[BCho]

Hey achille-

One of the reasons I asked to clarify what it is you wanted to measure is because of your curved images. Were you interested in the length of the particle if it were straightened? Or do simple want to find the horizontal distance (width) the particle takes on the screen? If you are interested in doing the later, The steps that Rialter pointed out would be very helpful. You could also use the particle analysis vi to determine the width of the particle (see attached image). With the example in the image you could find the width by doing a simple subtraction, or you could get a little more accurate and apply the distance formula to find the distance from the upper corner to the lower corner.

If you are more interested in finding the straight line distance from the tip of the particle to the root of the particle, you may want to consider using a caliper step to calculate the arc length of each curve and then add them together.

In terms of finding the width of the particle at the root (where it intersects the side of the image) you could use a find edge step to find the two edges at the side of the image and then use a caliper steps to determine the distance between the two points.

It would be helpful to take a look at what dimensions you want to measure and how you might measure that dimension in real life. Then, applying the tools in the Vision Development Module would be that much easier.

Hope this helps!

-Ben

Message Edited by BCho on 05-04-2009 10:15 AM