Skip to content
Intel RealSense R200 Gazebo ROS plugin and model
C++ CMake
Branch: master
Clone or download
Pull request Compare This branch is even with SyrianSpock:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
doc
include/realsense_gazebo_plugin
launch
models/realsense_camera
src
test
urdf
worlds
CMakeLists.txt
README.md
package.xml

README.md

Intel RealSense Gazebo ROS plugin and model

Important note

This branch only supports ROS Melodic with Gazebo 9.

To use this plugin with ROS Kinetic and Gazebo 7 you should use the kinetic-devel branch.

Quickstart

Build the plugin

catkin build realsense_gazebo_plugin

Test it by running

roslaunch realsense_gazebo_plugin realsense.launch

Run the unittests

After building the plugin, you can run the unittests

rostest realsense_gazebo_plugin realsense_streams.test

Run the point cloud demo

Using depth_image_proc package, we can generate a point cloud from the depth image by running

roslaunch realsense_gazebo_plugin realsense.launch # in terminal 1
roslaunch realsense_gazebo_plugin depth_proc.launch # in terminal 2

Then open Rviz, and display the /realsense/camera/depth_registered/points topic, you should see something like this Point cloud in Rviz

Run from URDF

roslaunch realsense_gazebo_plugin realsense_urdf.launch

This will behave the same as realsense.launch mentioned above, with the difference that it spawns the model from a URDF (see urdf folder). You can reuse this to plug the sensor in the robot of your choice.

Dependencies

This requires Gazebo 6 or higher and catkin tools for building.

The package has been tested on ROS melodic on Ubuntu 18.04 with Gazebo 9.

Acknowledgement

This is continuation of work done by guiccbr for Intel Corporation.

Thanks to Danfoa for contributing the URDF integration.

You can’t perform that action at this time.