Sunday, January 19, 2014

Machine vision a “coattails” industry?


Face it, for all the pronouncements about machine vision sales hitting $2bn, or $3bn, or even $6bn, ours is a pretty small industry. Complicating matters, developing new products, especially image sensors, is hugely expensive. Doing so for machine vision doesn’t make a lot of sense, which is why we’ve always made do with spinoffs from the consumer market.

I sense that’s changing. No, the vision industry is still a niche, but what’s going to drive technology development in the future are traffic applications.

If you want proof of this, log on to www.traffictechnologytoday.com and leaf through their magazine.

First, you’ll notice it’s thick, so there’s lots of advertising dollars around. Second, it’s stuffed full of adverts for camera-makers, companies that you’ve heard of, like Basler, Lumenera, Point Grey, and SVS-Vistek. Third, lens companies like Tamron are getting in on the action.

It’s interesting to speculate on what this means for us.

I see higher sensor resolutions in the future, plus greater dynamic range and improved ability to deal with temperature extremes. Likewise, increased sensitivity is important to deal with poor illumination. Pan-tilt-zoom is obviously big in the traffic world, though it’s never found a place in machine vision. And the same goes for zoom.

In terms of software, I would think traffic applications put more emphasis on dealing with fuzzy, unstructured or unpredictable images. They need to cope with wide variations in lighting and moving targets. There are probably other challenges I haven’t yet thought of too. How about color for instance? Is that relevant?

Overall then, I see this as being a good long-term development. Increased volumes for camera-makers should mean lower unit prices, and competition will also mean better products. It may even be that the traffic market becomes big enough to dictate sensor design, in which case we could have some very interesting products in our future.

Start your subscription to Traffic Technology magazine and watch my predictions come true.

No comments: