loading
 
Stereoscopic vision for determination of depth
Robert  [1 posts]
2 years
Hi
I am with the LEAF robotic group and am in the process of making an arm for my robot herbert. I want to try to use stereo vision to help detect where an object is located and be able to send those coordinates to herberts arm control software and circuitry. Currently the arm is built and works mechanically. I would appreciate any suggestions.

Bob
Steven Gentner from United States  [1371 posts] 2 years
Bob,

I'd recommend looking into the Kinect or Creative Senz3D instead which would give you that 3D information at a reasonable cost. Stereo is possible but you'll never get as good results as an active system such as these. The only reason you'd not use those is if you intend to operate in direct sunlight (i.e. outside). Stereo is more expensive since you need good calibration between the cameras in a stable rig. I've not see good Stereo pairs for less than the Kinect or Senz3D.

The trick to picking up things is to combine that depth information with the texture information to help segment the object. There is a lot of research in this area for example:

http://newsoffice.mit.edu/2015/algorithm-for-better-robot-object-recognition-0112

So you can find quite a bit on the web. Best to review and pick something easy to start with ... perhaps a task at home you'd prefer not to do?

STeven.

This forum thread has been closed due to inactivity (more than 4 months) or number of replies (more than 50 messages). Please start a New Post and enter a new forum thread with the appropriate title.

 New Post   Forum Index