Surface inspection

[skytalker]

Hi
I have the machine vision 's application about inspection the surface. The whole surface is divided into serveral sub-areas. The inspection performs by starting from on sub-area then moves to another. This application needs to ensure that when move from one point to another will not overlap to each other. For example, if starts from upper left conner and move to the right, the next sub-area will not be overlapped to the previous area. Please see attach file, illustrates inspection from one circle then move to another one-by- one.
I can acquire (live) image by using USB camera and the view of camera only see on one sub-area. I need suggesstion if I will move to another areas how can I control and avoid overlapping.
Regards,

[BruceAmmons]

1. Figure out the dimensions of your field of view.  You can put a scale on the surface, or use a dot pattern, or be creative.

2. If necessary, determine the motion scale to figure out how much the camera moves for a given motion.  You can make a small move and see how much a point moves in the field of view.  Since you have calculated the scale for the field of view, you should be able to measure the move in real world units.

3. When you move the camera, make sure the move is >= the field of view.

Does that answer your question?  It wasn't 100% clear what you are asking for.  Is each circle supposed to represent one field of view, or a small area in the field of view?  Are you trying to avoid inspecting each circle twice?

Bruce

[skytalker]

Hi ,
Yes it is , my target is to advoid inspecting one circle twice ( also they can not overlap to each other). One circle is represented as one field of view, when it is inspected then camera will move to another circle. I 'm trying to find the toolbox set which is can be used for the motion scalling task, but i cannot.
Regards,

[BruceAmmons]

There is no toolkit.  You have to do it the old fashioned way with simple math.

The first step is to figure out the scale of the image (units/pixel).  Put something with a known size or scale under the camera and measure it in pixels.  This gets you the total size of the field of view and the units/pixel.

The second step is to put an object with an easily trackable point on it.  Move the camera and measure the number of pixels between the old position and the new position.  You can use units/pixel to get the actual movement in units.

This works great as long as you have small amounts of distortion in your image.  As long as you aren't using a wide angle lens, you should be fine.

Bruce

[skytalker]

ah, I got the idea about this. The changing of the unit/pixel is sound nice. I 'm trying and let you know ,how does it go .
Regards,

[D60]

  Hi all!
First of all, Merry Christmas! 🙂 by the way, the work never ends. I am wondering about a solution, without any success. Could you help me?
My problem is the following. I have a surface, and I have to detect all of defects on it. I don't have any good idea how to find them. In the attachment, you can see a small surface defect. My project is: find and localize all of these defects automatically, and measure them. What kind of image processing tools can be used to find and measure (area) them?
Measuring them is not as important as than finding, and localizing them. 
First of all, using a morphological procedures could be a good start, but I don't know to detect these defect automatically...
Please see the attachment:

[muks]

---

durnek60 wrote:

  Hi all!
First of all, Merry Christmas! 🙂 by the way, the work never ends. I am wondering about a solution, without any success. Could you help me?
My problem is the following. I have a surface, and I have to detect all of defects on it. I don't have any good idea how to find them. In the attachment, you can see a small surface defect. My project is: find and localize all of these defects automatically, and measure them. What kind of image processing tools can be used to find and measure (area) them?
Measuring them is not as important as than finding, and localizing them. 
First of all, using a morphological procedures could be a good start, but I don't know to detect these defect automatically...
Please see the attachment:

---

What dont you start a new thread?

You have to enhance the image first using filters available.Then apply a proper threshold. do a particle filter and extract the defect. Attach a few images if possible.

[The_Chancenator]

Hi,

Please go ahead and start a new thread for this seperate question here. (Click on "Post New Thread"). 

Thanks!

[muks]

Continued here