A hidden benefit of machine vision

Has a “lean manufacturing” guru ever told you that vision systems don’t add value? I hear it a lot, and as you might expect, I take a different view.

If you adhere to the strict definition of “adding value” as a transformational process for which the customer will pay, then no, a vision system doesn’t actually help bring the input material closer to being finished. But if you were to use a broader definition, for example, “an activity that helps give the customer what he requires,” then vision adds value.

I can think of four ways in which this happens:

1. Identifying defective product – and ensuring it doesn’t get shipped.
2. Process guidance – locating a part correctly – a wheel on a hub, for example.
3. Process control – taking measurements and feeding the results to an upstream process.
4. Process improvement

I’ll bet that fourth point – process improvement – raised some eyebrows. How does machine vision help with process improvement?

This is the hidden benefit, and one that I wish people would make more use of. You see, vision systems are great at not just making measurements but at recording data. In other words, they can be a wonderful data acquisition tool, and data is the key to improvement.

Here’s what I recommend: look for ways to export key process data – dimensions, colors, whatever is important – to a database. For maximum value, make sure to timestamp every entry. Then, any time you want to understand the natural process variability it’s just a matter of querying the database. This way you can also identify the effect of any process changes you make – part of the old “Plan, Do, Check, Act” cycle.

So next time Mr. Lean Expert assails you, start by agreeing that machine vision doesn’t act upon the product being manufactured, and then ask if he believes in the importance of data in process improvement.

Let me know how you get on.