New Year resolutions

Around this time of year magazines, newspapers and local TV stations like to drone on and on about what resolutions we are all making, or perhaps should be making. You know the kind of thing … “this year I’m really going to exercise more,” we tell ourselves, or perhaps, “I really will lose ten pounds before my summer vacation.”

Well as it’s easier to swim with the current than against it, I too am going to share my (machine vision) resolutions.

Actually, there’s only one, but I have given it quite a bit of thought. So here’s my resolution: in 2011 I will spend much more time on educating my customer in what to expect from a machine vision project.

The reason I’m going with this, rather than for example, spend more time on lighting feasibility, or even, be nicer to the cat, is that I do have a tendency to oversell machine vision. When I’m trying to get the business I’m all, “yes we can do that, no it won’t be a problem, oh yes vision is proven technology,” and then I spend the next three months backtracking.

Obviously, that doesn’t sit too well with customers who have time and budget pressures of their own, so this year I’m going to be much more up-front about the risks and challenges. In particular, I want to:

1. Explain how a machine vision project progresses, and especially how much iteration he must expect to see.
2. Detail exactly what I need from him. Not just a dozen consecutive sample parts but boxes and boxes of product that will demonstrate the natural variability of the production process.
3. Educate him in why it’s not possible to have zero false rejects AND zero false accepts.

I realize this might cost me some business, but if I’m 100% honest, it’s probably business that I’m better off without, so I’m going to give it a go. Wish me luck!