Ask Your Question

lucasw's profile - activity

2019-10-09 05:14:47 -0500 marked best answer gazebo camera frame is inconsistent with rviz + opencv convention

It looks like the gazebo camera frame convention is not the same as rviz and opencv, which the image below shows

image description

In opencv, z is pointing into the image (the blue axis), x is right (the red axis), and y is down (green axis), while in the gazebo camera x is pointing into the image and z is up, y is right which is similar to the robot convention of x being forward and z up.

The image above is using an rviz/Camera to overlay the purple grid on the frame generated from the gazebo camera plugin, instead of the grid overlaying properly on the ground and going to toward the horizon rviz thinks the camera is pointed at the ground.

This example is running the gazebo_ros_demos rrbot_gazebo and rrbot_control launch files, and using standard Ubuntu 14.04 + Jade packages.

I cross posted https://github.com/ros-simulation/gaz... - or is it the fault of rviz/Camera and opencv, every node calling opencv camera projection functions should rotate first? Or every node on either side should have options to support either frame? (Or do options exist already and I've missed them?)

My short term solution is going to be to republish every frame out of gazebo with a rotated camera frame in the header (and the urdf/xacro can create the corrected frame, or it could be sent to tf from the same republishing node).

2019-10-09 05:14:47 -0500 received badge  Guru (source)
2019-10-09 05:14:47 -0500 received badge  Great Answer (source)
2019-10-08 09:02:17 -0500 received badge  Good Answer (source)
2019-10-06 09:47:58 -0500 edited answer rospy.msg.AnyMsg How do we use?

Here is a small example from @wkentaro that receives a messages, sleeps a certain amount and also adds that amount to th

2019-10-04 09:14:53 -0500 received badge  Famous Question (source)
2019-09-29 17:29:17 -0500 received badge  Great Answer (source)
2019-09-24 17:22:54 -0500 commented answer What are 'bond' topics for?

Is there an updated url for that ticket?

2019-09-24 14:15:19 -0500 received badge  Famous Question (source)
2019-09-07 08:42:37 -0500 answered a question 1D motion profile generation?

This is the closest I've found so far, but it relies on Matlab/Simulink exports to run: https://github.com/AIS-Bonn/opt

2019-09-05 16:27:58 -0500 commented answer Get timestamp during roslaunch

This almost works but there is an annoying trailing linebreak: <param name="dst" command="date +$(env ROS_WORKSPACE)

2019-09-05 16:27:49 -0500 commented answer Get timestamp during roslaunch

This almost works but there is an annoying trailing linebreak:

2019-09-05 16:27:28 -0500 commented answer Get timestamp during roslaunch

almost works but there is an annoying trailing linebreak.

2019-09-05 16:27:16 -0500 commented answer Get timestamp during roslaunch

<param name="dst" command="date +$(env ROS_WORKSPACE)%d-%m-%Y_%Ih%Mm%S.mp3" /> almost works but there is an ann

2019-09-05 16:27:00 -0500 commented answer Get timestamp during roslaunch

<param name="dst" command="date +$(env ROS_WORKSPACE)%d-%m-%Y_%Ih%Mm%S.mp3" /> almost works but there is an annoyi

2019-09-05 16:26:50 -0500 commented answer Get timestamp during roslaunch

<param name="dst" command="date +$(env ROS_WORKSPACE)%d-%m-%Y_%Ih%Mm%S.mp3" /> almost works but there is an annoyi

2019-09-05 14:46:27 -0500 edited question Eval roslaunch substitution expression from console

Eval roslaunch substitution expression from console Can I do that? For example $ roslaunch eval '$(find my_package)/so

2019-09-01 17:47:42 -0500 received badge  Self-Learner (source)
2019-08-24 23:57:23 -0500 received badge  Necromancer (source)
2019-08-22 13:28:56 -0500 edited answer rospy.msg.AnyMsg How do we use?

Here is a small example from @wkentaro that receives a messages, sleeps a certain amount and also adds that amount to th

2019-08-21 19:15:39 -0500 commented question ROS log file change of time-stamping format

It would be nice if were possible to use bash to process the console string, like put a date -d@${time} to convert the t

2019-08-21 19:12:19 -0500 edited answer ROS log file change of time-stamping format

After much digging around, I found this is possible, but the current logging architecture in ROS Kinetic makes it quite

2019-08-16 18:19:04 -0500 edited answer How to specify dependencies with "foo_msgs" catkin packages

I believe the modern way of doing this is to add a dependency on ${catkin_EXPORTED_TARGETS}, as specified on this docume

2019-08-14 22:28:35 -0500 received badge  Nice Answer (source)
2019-08-13 20:36:17 -0500 commented answer API to know if a node is alive?

The answer in #q164601 shows use of the rosnode api.

2019-08-09 20:52:59 -0500 received badge  Good Question (source)
2019-08-08 02:02:06 -0500 received badge  Famous Question (source)
2019-07-11 05:39:46 -0500 received badge  Famous Question (source)
2019-07-11 05:39:46 -0500 received badge  Notable Question (source)
2019-07-11 03:41:18 -0500 received badge  Famous Question (source)
2019-06-24 17:56:59 -0500 received badge  Good Answer (source)
2019-06-19 11:37:44 -0500 received badge  Notable Question (source)
2019-06-18 13:14:11 -0500 received badge  Self-Learner (source)
2019-05-30 17:23:19 -0500 edited answer Using rosbag to get size of each topic

I had the same issue and wrote a python script to solve it. #!/usr/bin/env python # Print total cumulative serialized

2019-05-20 02:12:06 -0500 marked best answer stdout to stdin pipe using launch-prefix?

I'd like to use launch-prefix to pipe the stdout of a program (netcat) to the stdin of a ros node.

It works fine when launched from rosrun:

netcat 10.0.0.10 8001 | rosrun foo stdin_to_msg.py

I can also copy the command line that roslaunch generates (determined by ps -aux | grep stdin_to_msg.py) and run that successfully, but when run in the launch file the stdout from netcat is sent to the stdout of the roslaunch terminal if output="screen" (or nowhere if not, or a log file?), and the node itself never starts.

<node name="stdin_to_msg" pkg="foo" type="stdin_to_msg.py"
    launch-prefix="netcat 10.0.0.10 8001 |" output="screen"/>

Probably I can wrap this in a bash script and run that from roslaunch, but I'm curious if there is an in-launch solution.

2019-05-13 05:42:30 -0500 received badge  Notable Question (source)
2019-05-12 09:42:34 -0500 received badge  Famous Question (source)
2019-05-10 20:27:29 -0500 received badge  Self-Learner (source)
2019-05-07 23:53:43 -0500 received badge  Nice Answer (source)
2019-05-04 21:31:14 -0500 received badge  Good Answer (source)
2019-05-02 18:31:08 -0500 commented question Accessing video frames of a rosviz display

It's controlled by the transform of the frame_id in the input camera_info header. You could broadcast that tf manually,

2019-04-22 21:45:09 -0500 commented question Using groups with dynamic_reconfigure, rqt_reconfigure

I changed the dynamic reconfigure request to only have one value in it (before it was resetting all the values when only

2019-04-22 01:08:23 -0500 received badge  Famous Question (source)
2019-04-21 21:47:39 -0500 commented question Using groups with dynamic_reconfigure, rqt_reconfigure

I'm getting the same error though only when the dynamic reconfigure server is a rospy one, an identical cfg with C++ ser

2019-04-21 21:32:31 -0500 commented question Using groups with dynamic_reconfigure, rqt_reconfigure

I'm getting the same error though only when the dynamic reconfigure server is a rospy one, an identical cfg from with C+

2019-04-19 10:41:29 -0500 commented question rqt plugin: crash with 'std::bad_alloc'

You should consider decoupling the data processing from the user interface in rqt and debugging would be simplified: hav

2019-04-15 15:04:59 -0500 commented question Accessing video frames of a rosviz display

Run source on your devel/setup.bash before running rviz. If you are using install/setup.bash it's possible I haven't ma

2019-04-15 10:11:42 -0500 commented question Caculating the distance of a point from point cloud data

The origin of the point cloud is likely near center of the sensor used to capture them- you could add detail about which

2019-04-15 10:09:27 -0500 commented question Caculating the distance of a point from point cloud data

The origin of the point cloud is likely near center of the sensor used to capture them- you could add detail about which

2019-04-15 10:07:05 -0500 commented question Caculating the distance of a point from point cloud data

The origin of the point cloud is likely near center of the sensor used to capture them- you could add detail about which