Defect identification and classification

[zulu]

Hi

Can anyone direct me the possibilities in labVIEW IMAQ Vision to find and classify the defects on the piece of sample attached.

[RManion]

Hello,

It looks like you are trying to do fabric analysis and search a texture for defects. There are several ways to do this within IMAQ Vision, but all methods will require implementing some algorithm or unique code in order to find the defect. Here are a couple of suggestions.

1) Start with VIsion Assistant. If you have a newer version of IMAQ Vision, the Vision Assistant program should be included. Vision Assistant is helpful for quickly getting started on projects like these. First import an image of your texture into Vision Assistant, and then perform various processing steps in order to find any defects. These processing steps may include anything from thresholding to morphology to pattern matching.

2) Also, I have seen many posts within these forums that are specific to fabric analysis. You might want to search the forum for 'fabric', 'defect' or texture. In addition, here is a link to a similar post.

Similar post

Additionally, Brian Beal has attached an image of the fabric after some basic processing. The link to your original post is here:

Original thread

3) Finally, a little more information regarding the nature of your application would be useful. Are you trying to classify particular defects, or just determine if any defect is present at all? Some of these cases are certainly easier than others.

These should be good places to start. If you have any questions feel free to let us know.


Robert

[zulu]

Hi Manion

My challenge is to classify the defects on fabrics of different texture,pattern and shades.Defects can occur randomly at any location.So am supposed to find out the kind of defect and its location on the sample
Am supposed to find out 60 defects defined.Am able to find out a few coarse defects using NI Vision asst.
But it couldnt find out most of the samples.
Can u help me out in this regard.

Zulu

[RManion]

Hi Zulu,

You mentioned in your post that you were able to find some of the 'coarse' defects, but you were not able to find all of the defects. If you are trying to find smaller defects that are present on many different fabric types, you will need a much more sophisticated algorithm for detection. I am not deeply familiar with all of these algorithms, but it is possible that you could find a suitable solution in a text on digital image analysis.

On a simpler level, I would try filtering and then thresholding, or perhaps some edge detection scheme. It becomes much more difficult to design an algorithm for detection when you expect variation within your valid input images. Maybe you could post a couple of the tougher images, and we could look at the ways to identify their corresponding defect?


Robert