Bees and ants are some of the favorite insects of roboticists as
evidenced by our frequent articles about them. So it should come as no
suprise that robotics researchers in Mexico are working on a stereoscopic
vision system inspired by the foraging methods of bees. It sounds a
bit bizarre but, after two cameras capture a scene in 3D, "virtual bees"
swarm over the image seeking features of potential interest based on
criteria such as texture and edges. "This technique consist[s] of
comput[ing] three-dimensional coordinates just for a few principal
points of the objects; after that, some computational entities -- the
bees -- will find the points between the principal ones, just like the
bees, in the nature, find flower patches from one flower." The areas
that get the most interest
from the virtual bees get the most attention from the computer when
rendering the image. The result is that the portions of the image most
important to robot navigation get rendered faster than less important
areas. Researchers Cesar
Puente and Gustavo
Olague of the CICESE Evolutionary
Vision lab hope to have "virtual bee vision" working on a
mobile robot by the end of 2006.
Yes i like the stufy of bees 3d vision etc. and their then guidance to
points of interest but surely this is study on a very limited
creatures abilities and is not the bees main guidance just their smell
system for honey etc ???,thank you though.
Steve,
This is the same algorithm as the one previously mentionned here:
http://robots.net/article/1695.html
except that the other one was called the fly algorithm. I imagine that
they improved it somehow.
Igor.
http://pegasusbridge.blogspot.com/
Although far more complex obviously, i beleive human eyesight is the
one for all people to work to duplicate,then make it even better than
ours.Might seem nieve but maybe a more important end result for the
world.
As far as I understand from reading this literature, the folks in France
are improving this algorithm want to apply it for blind people. So yes
Marev, it looks like they are tending toward duplicating the human eye.
Cheers IgorCarron,i have seen the work from france and some other
places on robotic human eyes,limited vision has already been acheived
and obviously once the initial vision is there it can be improved and
developed now for the future.I really do beleive in the importance and
the fact that the work hours should always be for curing and evolving
people not figuring out how a bug does it,cheers.