The site is read-only. Please transition to use Robotics Stack Exchange
| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
 | 1 | initial version |
1.) This was likely #defined as 3.141592 (pi) in one of the header files. Somewhere in ROS.
2.) I think the input cloud is defined in a camera frame. Camera frames are (annoyingly) different than the standard ROS frame. X is horizontal along the image, Y is vertical along the image (often facing down), and Z is positive going into the image plane. So here distance to point ahead is z and distance from center is x.
3.) No, it's not. However, if you are running these nodes on a slower computer, the throttling can help a lot. It's just a very timing-imprecise node to slow things down so that you don't have to process every pointcloud.
| | 2 | No.2 Revision |
1.) This M_PI was likely #defined as 3.141592 (pi) in one of the header files. Somewhere in ROS.
2.) I think the input cloud is defined in a camera frame. Camera frames are (annoyingly) different than the standard ROS frame. X is horizontal along the image, Y is vertical along the image (often facing down), and Z is positive going into the image plane. So here distance to point ahead is z and distance from center is x.
3.) No, it's not. However, if you are running these nodes on a slower computer, the throttling can help a lot. It's just a very timing-imprecise node to slow things down so that you don't have to process every pointcloud.
| | 3 | No.3 Revision |
1.) M_PI was likely #defined as 3.141592 (pi) in one of the header files. Somewhere in ROS.
2.) I think the input cloud is defined in a camera frame. Camera frames are (annoyingly) different than the standard ROS frame. X is horizontal along the image, Y is vertical along the image (often facing down), and Z is positive going into the image plane. So here distance to point ahead is z and distance from center is x.
In response to your edit, as stated in REP 103, the camera coordinate frame is "z forward, x right, y down", assuming you are looking from the image sensor to the image plane and are centered at the optical focus point of the camera.
3.) No, it's not. However, if you are running these nodes on a slower computer, the throttling can help a lot. It's just a very timing-imprecise node to slow things down so that you don't have to process every pointcloud.
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.
