The latest issue of IEEE Trans. on Robotics is all about SLAM. There are some interesting new algorithms, and articles on all the old ones. This stuff actually works now, with nothing more than a camera as input.
Willow Garage is implementing some SLAM algorithms and open sourcing the code by putting it back into OpenCV.
If your robot has at least PC-level compute power, there's no reason any more that it shouldn't know where it is and what's around it.
John Nagle
Do you think it's as good as - or approaching - Evolution Robotics vision recognition software ? What I mean is you could train the system by taking a photo against a monotonous background, and it would recognize the object later with a high degree of certainty.
Take a look at the SIFT and SURF algorithms for image recognition.
- Daniel
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.