registration of several reference images in a test image using vision builder

[RSFree101]

I have downloaded the trial version of vision builder, and I am attempting a test to determine if it's worth buying:

I would like to determine the position and orientation of two reference parts within a set of test images.  I am able to load the reference images, obtain the grayscale, perform appropriate filters, determine their global orientation based on some straight lines, and select the desired ROI's that I would like to match up.  I am also able to load in a test image, take the grayscale, and filter it.  My state diagram has three heads with the three loading steps that converge. This all seems to be great, but how do I find the position and orientation of the reference ROI's within the test image? 

I would like to use something like the geometric matching tool, but it only seems to work on using a part of a first image on subsequent images in the same folder.  I suppose I could use my reference as the first image in the folder, then loop through the remaining ones, but...

Once I find the position and orientation of the first reference in the test image, I would then like to find the position and orientation of the second reference in the same test image, then report values like relative distances and angles between these found references.

Does this mean I need two folders, with their own reference image and the same set of test images?  Would I combine the geometric matching results using some variables or math functions?

What is the best way to perform this image registration in vision builder?

[Moorethoughts]

RSFree101,

Thank you for your interest in Vision Builder AI.  If I understand what you are trying to do, you are looking to compare the position locations of two separate parts in an image.  We have a tool that does this very well.  If you use the pattern matching to find both instances of the part, you should be able to determine the distance and angle between them relatively easily.  To do this, you will need to find an edge in between the two so that you can justify the angle between them, and then simply use a few Geometry functions to actually find the angle.  To record this information, you can simply use a Data Logging step to write your findings to file.  

If you are looking for a more complete analysis on how you might be able to do this, you will need the Vision Assistant on your computer (if you do not have it already, you can get it by downloading a free trial of our Vision Development Module).  In the Vision Assistant splash screen, click on the Solution Wizard.  This pulls up a menu.  From there, go to Tutorial » Bracket Inspection.  Click Load Solution.  If you step through that example, this displays how to do what you are looking to do with the Vision Assistant, but it should be no different at all in Vision Builder AI.  Hopefully this helps.

Regards,

Keith M.

Applications Engineer