[ros-users] camera1394 and swissranger

Blaise Gassend blaise at willowgarage.com
Tue Mar 30 16:45:52 UTC 2010


The calibration would be done at a higher level, and the parameters
passed back to the camera driver whose responsibility it is to publish
them. So if you just output default values, people will have a hard time
using the device as a camera.

You could look at the CameraInfoManager in camera1394. It would be easy
for you to pull into your driver, and would allow people to calibrate
the camera just like any other camera. At some point I want to move that
functionality into some common package, but I haven't had a chance yet.

Blaise

On Tue, 2010-03-30 at 11:12 -0500, Patrick Beeson wrote:
> The swissranger is calibrated at the factory.  They do not provide the 
> internal optical parameters they use when getting 3D points from 2D 
> images.  If one wanted to rectify the 2D depth or intensity images, one 
> could run a checkerboard calibration from the intensity image (I've done 
> this).  I'm assuming this would be done at a higher level than the 
> camera communication driver.   So, for now, camera parameters will need 
> to be set to defaults.
> 
> 
> Blaise Gassend wrote:
> > If it looks like a dog, and barks like a dog, why not put it in
> > dog_experimental?
> > 
> > So if you are publishing an image and a camera info with the standard
> > directory structure, it sounds like a good candidate for
> > camera_drivers_experimental.
> > 
> > How are you dealing with calibration?
> > 
> > Blaise
> > 
> > On Tue, 2010-03-30 at 09:44 -0500, Patrick Beeson wrote:
> >> I think I'll have some time later this week to update my ROS Swissranger 
> >> 3000/4000 driver with appropriate image_proc info.  Radu and I had 
> >> discussed this replacing his older 3K node, as this is basically a port 
> >> of that to work with the newer Mesa Imaging API.
> >>
> >> I think having this driver in the camera_drivers_experimental would be 
> >> fine, as the device is essentially a camera that publishes multiple 2D 
> >> images along with a point cloud (I'll make this a runtime option).  Any 
> >> objections to locating a 3D flash lidar driver in the 
> >> camera_drivers_experimental stack?
> >>
> >>
> >> Patrick Beeson
> >> _______________________________________________
> >> ros-users mailing list
> >> ros-users at code.ros.org
> >> https://code.ros.org/mailman/listinfo/ros-users
> > 
> > 
> > _______________________________________________
> > ros-users mailing list
> > ros-users at code.ros.org
> > https://code.ros.org/mailman/listinfo/ros-users
> _______________________________________________
> ros-users mailing list
> ros-users at code.ros.org
> https://code.ros.org/mailman/listinfo/ros-users





More information about the ros-users mailing list