Errors in camera_pose_calibration launch for 3 Kinects
So after finding no alternatives in RosHydro, I installed the Fuerte version of camera_pose_calibration and added the fuerte stacks/share to my ROS_PACKAGE_PATH. However, when I try to launch it I get a lot of errors. The thing I find interesting about these errors is my printout remains the same regardless of whether or not I've currently got the three Kinects launched. This appears for all 3:
ERROR: cannot launch node of type [image_cb_detector/image_cb_detector_action_old]: can't locate node [image_cb_detector_action_old] in package [image_cb_detector]
Then this:
[capture_exec-20] process has died [pid 22968, exit code 1, cmd /opt/ros/fuerte/stacks/camera_pose/camera_pose_calibration/src/camera_pose_calibration/multicam_capture_exec.py kinect_1 kinect_2 kinect_3 request_interval:=interval_filtered __name:=capture_exec __log:=/home/robot2/.ros/log/dc0c0b7e-7709-11e3-8620-b8ca3a789964/capture_exec-20.log]. log file: /home/robot2/.ros/log/dc0c0b7e-7709-11e3-8620-b8ca3a789964/capture_exec-20*.log
Then this repeating while I have it running:
[ERROR] [1389037001.654864018]: Unable to convert '8UC3' image to bgr8: '[8UC3] is not a color format. but [bgr8] is. The conversion does not make sense'
Any ideas? I'd gladly use an alternate package if one existed.
Edit: I also received a warning when I tried to launch calibration_tf_publisher.launch:
[ WARN] [1389039127.099446312]: You've passed in true for auto_start for the C++ action server at [/tf2_buffer_server]. You should always pass in false to avoid race conditions. process[cal_publisher-4]: started with pid [23549] Topic Name: camera_calibration Couldn't find file [/tmp/camera_pose_calibration_cache.bag]. Skipping publishing
>The "can't locate node [image_cb_detector_action_old" goes away using the camera_pose stack from github.com/jbohren-forks/camera_pose.git" Is this a pretty typical solution in the community or one that you discovered? I haven't seen this mentioned before.
So you never got multiple Kinect views pieced together in RViz at all?
Interestingly enough, now I'm not getting the 8UC3 error when I try to run it through groovy. But I'm still getting two others.
So just a heads up, I got no 8UC3 image while running with Hydro right now. I launched my rgb/raw/disparity windows first for all 3 Kinects. Then I sourced Fuerte in a new terminal (not in .bashrc) and launched it and got 3 small fuzzy windows in the aggregated image. It's a start.