Thursday, August 14, 2008

Learning from experience

I fell over a lot when I was learning to skate, and sometimes it hurt. But that’s how we learn; by getting it wrong, adjusting and trying again until we get it right.

Machine vision projects are no different; the wise engineer learns from his mistakes. The problem is that whereas I could keep falling over for as long as I wanted to be on the ice, few companies will allow an engineer to make repeated mistakes. Indeed, a single mistake can be career-limiting, especially if it’s a big one.

So how do you learn about the application of machine vision technology in a risk-free way? One method is to study what other people have done. The trade press is full of case studies, often written by the vendors of vision technology. (Cognex and National Instruments are particularly good at this.) For my money,
Vision Systems Design magazine is one of the best places to go for this type of article.

But when did you last read a case study about a vision project that failed? I guess it’s human nature to brush our mistakes under the rug, but that’s a pity because I feel sure we could all learn more by analyzing what went wrong than by admiring a successful implementation.

So let’s hear about the disasters. Who wants to go first?

No comments: