3D Robot Vision based on Stereo Vision

The task is to pick object from a box with a robot.
Scorpion 3D Robot Vision is used to solve the task.
The images shows the object captured with the two cameras.

stereo_picking

The specification is as follows:

  • Pixel resolution: 1.5 mm
  • Pallet size: 1200 x 800 mm
  • Number of layers: 6
  • Z- span of layers are from 550 mm
  • Z distance to Camera from 3D base reference system/robot work object: 2490 mm
  • Distance btw cameras: 300 mm
  • 3D calibration rms value: 0,3
  • Two Prosilica GC1020 GigE camera
  • Scorpion Vision Standard with 3D option version 6.3
  • Industrial PC – Intel Quad-Core 2.4 GHz
  • Windows XP SP2

3D Stereo Vision Height Profile

Captured Image:

  • This image shows that the six layers spans 350 mm height and that
  • the height difference within one layers is 15 mm.
  • The accuracy in z is plus / minus 2-3 mm.

5 Responses

  1. [...] in Class accuracy and robustness with Stereo Vision with up to 4 [...]

  2. [...] information: 3D Robot Vision based on Stereo Vision Possibly related posts: (automatically generated)Sony Scorpion Robot InspectionImportant [...]

  3. [...] More information: 3D Robot Vision based on Stereo Vision [...]

  4. [...] A classic example of how 3D machine vision can resolve these kinds of problems, this paper presents the problem and discusses the solution. More information TordivelBlog. [...]

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 48 other followers