MSA on Auto Vision Inspection (pass/fail) - No operators


Hello everyone,

I'm currently working on setting up an MSA study for an automated vision inspection system. This system relies on a camera to evaluate the positioning of strips on a plastic cassette, determining whether they meet the pass/fail criteria (attribute data). The overall process involves assembling cassettes, cutting LFI cards into strips, and passing through the vision inspection which identifies issues like incorrect positioning, wrong orientation, and strip damage.

Given that the system operates automatically, there's no human interaction involved.

I'm contemplating an Attribute Agreement Analysis using Minitab. However, I'm uncertain about how to proceed since I don't have any appraisers involved.
I was thinking on selecting 30 samples, a mix of both acceptable and defective cases. Then, do 5 randomized replicates to assess repeatability. I would only have 1 appraiser I believe because I am the one creating the samples and feeding into the system. Is this okay? Note that the only interaction of the operator with the system is the loading of the raw materials, which does not impact the vision test or cassette assembly process at all.

Is there any MSA analysis in Minitab that can be done for this particular case?

I eagerly welcome insights and recommendations from experts in this field as I am a little lost on what approach makes the most sense for this case.

Thank you!


Forum Moderator
Minitab has an Attribute Agreement Analysis that will handle a single appraiser. You typically look for a percent agreement > 90% and
Fleiss' kappa > 0.8 per table below.
Top Bottom