Image Matching with Labview Golden Template

[asapount]

Hello everyone!

I am currently trying to implement an image matching algorithm on two .png images (let's call them Template and Photo).

The idea is that Template contains an examply of what should be shown in a certain part of the Photo.

The Template and Photo have different image sizes.

What we want to achieve is :

1.Find the appropriate ROI in the Photo, where the Template is likely to be found

2. Compare the ROI with the actual Template

3. Get a similarity percentage as result (e.g. >90% similar -> Definite Match , 70-89% -> Possible Match , <70%similar -> Not a match)

Do you have any ideas of how we could implement this?

We are currently using LABView 2015 SP1 - 32bit with the Vision Add-On

We have considered using a modified version of the Golden Template Inspection.vi (can be found) but are stuck because the IMAQ Compare Golden Template Module keeps on returning error message attached to the post.

Any feedback on possible debugging is also welcome!

Thanks!

[Bob_Schor]

Did you read the Golden Template Comparison "Concepts" section in NI Vision Concepts Help?  [Look under Machine Vision].  This is not something I do, myself, so I hope you can "take it from here" ...

Bob Schor

[udka · Accepted solution]

A template is *.png with vision info (for e.g., matching parameters, info etc). You have to create the template using template editor which should be installed and accessed from start menu.

[asapount]

The Golden Comparison turned out to be more useful for showing the results, rather than obtaining them.

Using the Geometrical Pattern and Feature Pattern recognition with the Template editor gave us the first similarity score and expected coordinates/scaling/rotation.

Afterwards we applied the golden pattern to show the differences on the picture.