shape recognition opencv


I have to admit to struggling a little here

If someone could lend a hand it would be appreciated

I would like of to recognize particular shapes in a live video image eg a cross

I have got to the point where I have written a line of code to draw a box at the start of each blob using ofcontourfinder but i can’t seem to progress further

Am on xp with codeblocks

I am more used to Processing

with thanks



i don’t want to be too discouraging, but recognising particular shapes is very, very difficult. the best current methods (in the research literature) tend to involve training bayesian filter trees. there’s lots of research out there, and i’m sure it’s possible, but you’ll probably need to start reading academic journal articles.


the oreilly opencv book has a section about matching contours:

might be a good starting point for seeing ways to do it.

  • z