Log in or register to post comments

Vuforia for Defect Detection

November 5, 2018 - 5:02am #1

Greetings everyone, I've started working on a project for defect detection and I would like to know if Vuforia and its object tracking tools are capable of accomplishing this. So far I was just testing it with a binary case. I would like to know if, for example, a cube is perfectly tracked, it will to return a "positive" output but if I remove a corner from it or deform it in any way it would return a "negative" output(it still recognizes the object but its not completely the object it was looking for). I dont know if you can get the % of data points that are getting matched with the object scanned...

 

 

I can somehow achieve this using a trained neural network model, but since this is just a prototype I'd really be happy if vuforia could make this process easier.

 

 

Thank you, have a great day.

Vuforia for Defect Detection

November 12, 2018 - 6:22am #3

Thank you for the answer, but is there any way to get some kind of how many data points are being matched by the object that is being scanned? For example if I put a finger on a scanned object, it will still detect it.

 

But perhaps you are right and this is not the best use case for Vuforia.

Vuforia for Defect Detection

November 6, 2018 - 1:06pm #2

Hello,

I don't think that this is a good use case for Vuforia.

The Object Targets and Model Targets features are optimized to detect and track the recognized object. If the object has been altered enough such that Vuforia cannot detect it, then you will receive no information from Vuforia other than it is not being tracked even though the altered object's in the camera's field of view.

Thanks,

Vuforia Engine Support

Log in or register to post comments