Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Multi-camera synchronization, IR interference, and USB bandwidth questions #138

Closed
Thanh-Binh opened this issue May 11, 2016 · 11 comments
Closed
Labels

Comments

@Thanh-Binh
Copy link

I have some questions by using multi-camera:

  1. How can we synchronize all cameras?
  2. how is the video frame rate if all camera connect to the same USB3-port / to different USB3-port?
  3. How is the interference between cameras if they are oriented to the same object?

It is very nice to hear any answer from your experience!
Best thanks

@mhkabir
Copy link

mhkabir commented May 11, 2016

Same questions! :)

@dorodnic
Copy link
Contributor

  1. Depends on what you mean by synchronization. Temporal sync (making sure all cameras capture frames at the same time) is problematic since each camera has its distinct clock. For spacial sync, you could match objects from one camera in frames of another through image registration or estimate fixed relative camera models. Both techniques are not trivial, and are beyond the scope of librealsense right now.
  2. Using the same USB3 controller for more then one realsense device is not recommended. Bandwidth of USB3 is limited and you will inevitably loose frames. If you're using separate USB controllers, you will still need a powerful PC or you will have frame drops at the kernel driver level.
  3. The interference depends on camera type. In short, you can capture depth from two R200 devices pointed at the same target, but not from two SR300/F200. You can read more about that under Camera Specs

Perhaps @ddiakopoulos will be able provide additional information.

@Thanh-Binh
Copy link
Author

I mean temporal sync for object tracking and 3D pose estimation.
Best thanks

@bloyl
Copy link

bloyl commented May 12, 2016

@dorodnic
regarding interference from multiple SR300/F200 cameras aimed at the same target do you know if anyone has used the shake and sense method to alleviate this problem? Is there any reason to think that it would work less well on the SR300 and F200 then on the kinect V1?

@ddiakopoulos
Copy link
Contributor

@bloyl Kinect V1 uses very traditional structured light (projected dot pattern). The SR300 and F200 both use a "coded light" approach that computes depth over a series of 11 grey-coded frames -- this means it is very sensitive to motion.

@bloyl
Copy link

bloyl commented May 12, 2016

Thanks, that makes sense.

Are there any issues with mixing R200 cameras with an SR300 or F200? does the R200's laser cause an issue to the "coded light" depth sensing.

How does the depth resolution of the R200 camera, used in the 0.5m-1m range match up to the F200's depth resolution?

also unrelatedly, is there a better place to ask these questions then on the bug tracker?

@teknotus
Copy link

Are there any issues with mixing R200 cameras with an SR300 or F200? does
the R200's laser cause an issue to the "coded light" depth sensing.

Actually the opposite. If an F200 is pointing at the same spot as an R200
the depth data from the F200 is mostly intact, but the R200 data is pretty
much ruined.

@Thanh-Binh
Copy link
Author

@dorodnic
regarding the temporal synchronize between many cameras:
Is it possible to trigger camera from SW, because each camera is global shutter?

@ddiakopoulos
Copy link
Contributor

Hi @Thanh-Binh unfortunately not. Current-generation RealSense cameras are not designed to compete with the feature set of industrial vision sensors.

@digitologist
Copy link

Have you checked the infrared viewer on the r200 when this happens? Likely simple oversaturation and reducing the IR gain or exposure would bring it back.

On May 13, 2016, at 5:19 AM, Daniel Patrick Johnson notifications@github.com wrote:

Are there any issues with mixing R200 cameras with an SR300 or F200? does
the R200's laser cause an issue to the "coded light" depth sensing.

Actually the opposite. If an F200 is pointing at the same spot as an R200
the depth data from the F200 is mostly intact, but the R200 data is pretty
much ruined.

You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub

@ddiakopoulos ddiakopoulos changed the title different questions on multi-camera Multi-camera synchronization, IR interference, and USB bandwidth questions May 20, 2016
@RocketSumit
Copy link

I am using one USB3 controller for 4 SR300 cameras and losing frames for RGB images too.!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

8 participants