| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
Is there a way to calculate or extract the camera location and pose within RGBDSlam. I've been poking at the RGBDSlam code for a while trying to figure out if data was able to be published or calculated somewhere but can't find anything. I think it is inferred somehow as the newest version prints the changed coordinate axis on each update. Thanks!
In the new snapshot (see http://www.ros.org/wiki/rgbdslam/evaluation) you can store the trajectory to file. Also, if you send the model, a transformation from /map to /openni_camera is published. In the svn version, a transformation from /openni_camera to /slam_transform is published, where (somewhat confusing) /openni_camera acts as the fixed/map frame and /slam_transform is the moving camera postion.
Not sure, but in GraphManager.cpp there is a method publishCorrectedTransforms(). Maybe it has what you are looking for. Otherwise you should check the addNode() method of GraphManager. I think that is where all the magic happens.
Asked: 2011-11-01 01:50:06 -0500
Seen: 671 times
Last updated: Nov 02 '11
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
