How do you improve beginner tips when working with Robot Vision?

[falconjade]

I have a question about Robot Vision that may be useful for other people working with it too.

I am particularly interested in beginner tips and learning curve. I would like to compare approaches that people have actually used rather than just theoretical suggestions.

What has worked best for you, and what would you avoid?