Getting camera parameters

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Getting camera parameters

Robin
Hi all,

I'm writing some ROS code that needs to access the RGB and IR camera intrinsics. I thought I could do this by subscribing to the messages published on the topics /camera/rgb/camera_info and /camera/depth/camera_info, but the matrices (D, K, R, P) I get from those appear to identical, which doesn't seem right.

Is there a better way to obtain the intrinsics through msgs? I saw a post elsewhere about using the command GetFieldofView in the DepthGenerator class, but I don't know how to setup my code to do that.

Best
~Robin
Reply | Threaded
Open this post in threaded view
|

Re: Getting camera parameters

Suat Gedikli
Hi,

the camera parameters for the depth stream depend on whether registration is turned on or off. If the registration is turned on, depth images are mapped to the RGB frame, thus need to have the same camera parameters. So, if you turn of the registration, you should see the change in the parameters.

cheers,
Suat Gedikli
Suat
Reply | Threaded
Open this post in threaded view
|

Re: [Ros-kinect] Getting camera parameters

Hordur Johannsson
On Wed, 2011-03-02 at 15:59 -0800, Suat Gedikli wrote:
> Hi,
>
> the camera parameters for the depth stream depend on whether registration is
> turned on or off. If the registration is turned on, depth images are mapped
> to the RGB frame, thus need to have the same camera parameters. So, if you
> turn of the registration, you should see the change in the parameters.
One thing I noticed with that is that when the images are registered the
focal length used is a constant, rgb_focal_length_SXGA_, while when they
are not registered depth_focal_length_SXGA is used. Which seems to be
read from the camera (see openni_device.h, openni_device.cpp). Is that
correct? Shouldn't the rgb_focal_length_ be a calibrated focal length?

Thanks,
Hordur


_______________________________________________
Ros-kinect mailing list
[hidden email]
https://code.ros.org/mailman/listinfo/ros-kinect