Some Thoughts on Machine Vision

We are getting a lot of questions about machine vision and, yes, we have machine vision capable printers. We are not promoting this feature yet for a few reasons.

We do not feel that machine vision is not yet where it needs to be to support production scale. I have seen this process in person and first impression is - it's damn cool. Just drop a bunch of items on the print bed and the system automatically figures out how to orient the art and apply it to the objects. Outstanding.

When you really dig into how this would fit into a larger scale production, however, there are some challenges;
1. The cost is still high - in some cases 25%-40% of the actual system price.
2. Print registration is very good, but not at the level of accuracy and repeatability you would achieve with typical template-based operations.
3. Requires a heavy premium of setup on initiation and on a per-job basis, as well as regular maintenance and calibration that is not necessary with standard operations.
3. It does not optimize the space on the bed, nor does it play well with robotic loading operations.
4. Difficult to consolidate multiple customer orders reliably and mange them consistently through to pack and ship.

The technology is developing quickly and none of the issues above are insurmountable. In fact, the tech could be appropriate now for the right niche application. Companies looking to scale production must carefully weigh the benefits of automation against the realities of system complexity and operational needs. For now, we will continue to work on developing this with a mind toward fit for purpose solutions.

Have you encountered challenges with machine vision in your printing processes - or do you have a different opinion? We'd love to hear your thoughts and experiences.