0
$\begingroup$

Rosanswers logo

Hi all,

I am having some problems with this tutorial: http://www.ros.org/wiki/openni_launch/Tutorials/ExtrinsicCalibration

I calibrated both the RGB and the Depth camera on the Kinect, however, when I try to do this tutorial. I always get an error which says "Timed out waiting for checkerboard". I tried to find this error, however, I don't see anyone else having this problem.

I'm not sure if it is because the tutorial is broken for ROS fuerte. Anyone else having problems getting this to work?

Or if there is a better calibration method out there on ROS to calibrate the kinect. Note: I'm trying to get a better calibration than the manufacture. This is because I need more accuracy (millimeter range) than what is given (centimeter range).

EDIT: I tried to use the Contrast Augmentor using this command:

rosrun contrast contrast_augmenter image:=/camera/ir/image_raw

Then the input to the calibration was:

roslaunch camera_pose_calibration calibrate_2_camera.launch camera1_ns:=/camera/rgb_bag camera2_ns:=/camera/ir_augmented

Its the standard checkerboard so I should not need to give rows and columns. I did not have to when I did the calibration for the RGB Camera and IR Camera.

I still get the error on the screen "Timed out waiting for checkerboard". Am I missing something here?

EDIT_2:

I tried adding in the Checkerboard size and the number of columns and rows. Still can't be seen.

EDIT_3: image:


Originally posted by astaranowicz on ROS Answers with karma: 238 on 2013-05-15

Post score: 0


Original comments

Comment by jcc on 2013-05-23:
on both RGB and IR? What version of ROS are you using?

Comment by astaranowicz on 2013-05-23:
I'm using fuerte. I'm only running the Contrast Augmentor on the IR image. Also I'm getting the error message on both the RGB and IR image.

Comment by jcc on 2013-05-26:
strange. Yes, the Contrast Augmentor is only for IR. This packaged seams dead and without a substitute. I tried to use it on groovy (compiled from the source) and there is a error with cv_bridge on a conversion.

Comment by astaranowicz on 2013-05-28:
Well if the package is dead. I was able to get the intrinsics. Still need the extrinsic though.

Comment by CHz on 2013-05-30:
Can you try to put checkerboard more closer and horizontal. It seemed a bit small to me.

Comment by astaranowicz on 2013-05-30:
That was just one instance. When I calibrated separately, that seemed like the best and stable position in both cameras. I moved it closer and change its position, still the same.

$\endgroup$

2 Answers 2

0
$\begingroup$

Rosanswers logo

Try using this: contrast-augmenter

The theory is that there is a bug somewhere on the image pipeline (there was a topic explaining that), if i'm not mistaken the imageshow is using the 8 LSB and the checkerboard is using the 8 MSB, so you need to shift the bits to transform the LSB in MSB. The node that i liked does that for you.

Don't forget to make that the new input to the camera calibration program


Originally posted by jcc with karma: 200 on 2013-05-19

This answer was NOT ACCEPTED on the original site

Post score: 1


Original comments

Comment by astaranowicz on 2013-05-20:
Thank you for the tip. I'll try this out and report the results. I think the tutorial should be updated if this is a common problem.

$\endgroup$
0
$\begingroup$

Rosanswers logo

The problem was solved by reading this website: Link.

The most important part from this sequence was this line: " At the time of writing, the camera_pose_calibration package doesn’t work for us. There is a patched version that can be gotten from: https://github.com/jbohren-forks/camera_pose.git "

For some reason, the code-base on git-hub has some fixes which are not included in the version on ROS.


Originally posted by astaranowicz with karma: 238 on 2013-05-30

This answer was ACCEPTED on the original site

Post score: 1


Original comments

Comment by Athoesen on 2014-01-09:
Is this still the case, that it is not updated?

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.