Task 1 Camera Frames

[jiang-yiwei]

Hi,

In task 1, can I assume that the offset between the left camera and the right camera is a known constant?

By looking into the yaml file/subscribing to the topics, I see the transformation is only a 0.04 translation offset in the horizontal direction. In other words, is the optical center of the cameraR always at (0, 0.04, 0) in cameraL's frame?

Thank you so much!

Yiwei

[adnanmunawar]

Hello Yiwei,
Yes, the camera offset is going to remain constant throughout the 3 tasks. And that is correct, the cameraL is 0.02 Simulation Units left of the CameraFrame and cameraR is 0.02 Simulation Units right of it in the y-direction.

[jiang-yiwei]

Thank you for the confirmation!

[kschwan]

Hi guys,
Can I ask a related question --

Are the stereo camera's optical axes supposed to be parallel? They are not currently. Cameras are translated wrt. each other, but they "look at" the same point, i.e.,

cameraL:
  location: {x: -0.02, y: 0.0, z: -0.5}
  look at: {x: 0.0, y: 0.0, z: -1.0}
cameraR:
  location: {x: 0.02, y: 0.0, z: -0.5}
  look at: {x: 0.0, y: 0.0, z: -1.0}

Visualized:

Instead using the following parameters, the optical axes become parallel:

cameraL:
  location: {x: -0.02, y: 0.0, z: -0.5}
  look at: {x: -0.02, y: 0.0, z: -1.0}
cameraR:
  location: {x: 0.02, y: 0.0, z: -0.5}
  look at: {x: 0.02, y: 0.0, z: -1.0}

[adnanmunawar]

Hi Kim,

Yes, the two cameras are currently in a toe-in configuration. Do you think this makes the tasks more difficult?

[kschwan]

No, not really. I was just curious to know whether it was intentional :)

[adnanmunawar]

Well glad to hear that 😊. We just chose the toe in configuration for no particular reason other than maybe it giving better view on the DVRK stereo console.