Adrian Jongenelen, a PhD student at Victoria University, and his supervisor Professor Dale Carnegie explain their work in developing a new type of eye for robots that will allow them to ‘see’ in 3D. They will be able to see the shape of objects and even ‘recognise’ faces. Unlike human eyes that see things in 3D instantly, robot eyes need to scan every part of the surface to see it.
Point of interest
What other uses do these robot eyes have?
I'm dealing with the robot’s eyes. We want to make the robot be able to see where it’s moving and so it doesn't bump into objects or if it sees something coming at it, it can try to avoid it or react to it.
The current sensors that MARVIN is equipped with, he can sort of see point sources. He will shine a laser beam in a single direction, and he will know what there is in that one direction, whereas, what I'm working on, I want to get a whole 3-dimensional image of the scene. He will be able to actually get an idea of the shape of objects, and I guess he could also look at people, get the idea of a shape of someone's face, maybe even recognise them.
PROF DALE CARNEGIE
If we want to map a corridor or a room, we have to scan that laser dot or scan that infrared beam over every single part of that surface, and that can take quite a bit of time. And that is where our eyes are great – we can do it all at the same time, everything that we can see as opposed to just that one point.
The immediate application here at Victoria would be to mount this on mobile robots so give them better vision of the environment they are in so they can avoid things or recognise things. Also could use the sensor in quality control, so say a manufacturing line, looking at parts, making sure that they are the right shape.