realsense_ros2

ROS 2 wrapper targeted for the Intel RealSense T265 tracking camera on ROS 2 Jazzy.

This wrapper's implementation is specially developed with the objective of running it in Nvidia's Jetson Nano, however it should also work on any other platform running Ubuntu 18.04 and 20.04.

By running this wrapper you would be able to obtain:

• Pose data from the realsense t265 tracking camera

Legacy/optional support in this repository also includes D435 depth camera nodes.

Tested on Jetson Nano: L4T 32.4.3 [ JetPack 4.4 ] Ubuntu 18.04.4 LTS Kernel Version: 4.9.140-tegra ROS 2 Eloquent

Tested on Ubuntu 20.04 amd64: ROS2 Foxy

Requirements

• ROS2 jazzy
• Ubuntu 24.04

Installation

1. Install librealsense2 of the version 2.50.0 from sources:

   # dependencies
   sudo apt update && sudo apt upgrade
   sudo apt install git wget cmake build-essential libssl-dev libusb-1.0-0-dev libudev-dev pkg-config libgtk-3-dev libglfw3-dev libgl1-mesa-dev libglu1-mesa-dev
   
   # download librealsense2 version 2.50.0
   git clone https://github.com/realsenseai/librealsense
   cd librealsense
   git checkout v2.50.0 
   
   # build and install
   mkdir build && cd build
   cmake .. -DFORCE_RSUSB_BACKEND=ON -DBUILD_EXAMPLES=true -DBUILD_WITH_CUDA=OFF
   make -j$(nproc)
   sudo make install

2. Connect your cameras and check they are working propperly by openning a new terminal and typing:

   realsense-viewer

   You should be able to add the t265 and d435 cameras and view their data streams simultaneously.

   

3. Create a ROS 2 workspace, clone this repository inside and build:

   mkdir -p dev_ws/src
   cd dev_ws/src
   git clone https://github.com/jdgalviss/realsense_ros2.git
   cd ..
   colcon build

4. Source your workspace

   . install/setup.bash

Run

T265 tracking camera (primary target)

Run the node with the following command:

ros2 run realsense_ros2 rs_t265_node

Optional T265 parameters:

• enable_pose_jumping (default: false): Enable/disable T265 pose-jump handling.
• enable_relocalization (default: false): Enable/disable T265 relocalization.

Fisheye TF correction is applied internally as a fixed rotation (roll +90 and yaw -90).

To also publish fisheye images from T265, run:

ros2 run realsense_ros2 rs_t265_node --ros-args -p publish_fisheye:=true

When publish_fisheye is enabled, static transforms from t265_frame to t265_fisheye1_frame and t265_fisheye2_frame are also published using RealSense API extrinsics.

D435 depth camera only

Run the node with the following command:

ros2 run realsense_ros2 rs_d435_node --ros-args -p is_color:=true -p publish_depth:=true -p fps:=30

• When the is_color parameter is set to true, the color image from the depth camera is aligned with the pointcloud.
• When the publish_depth parameter is set to true, the depth image is published.
• When the publish_pointcloud parameter is set to true, the 3D pointcloud is published.
• The fps parameter is used to modify the rate at which the node publishes the pointcloud and the depth image.

T265 tracking and D435 depth cameras simultaneously

In one terminal, launch the two cameras:

ros2 launch realsense_ros2 realsense_launch.py

To launch cameras with RViz2:

ros2 launch realsense_ros2 realsense_rviz_launch.py

T265 tracking and D435 depth cameras simultaneously with SLAM-TOOLBOX 2D SLAM

In one terminal, launch the 2 cameras:

ros2 launch realsense_ros2 realsense_2d_slam_launch.py

In another terminal launch slam-toolbox:

ros2 launch realsense_ros2 online_async_launch.py

This requires to have ROS Slam-toolbox installed (apt-get install ros--slam-toolbox)

T265 tracking and D435 depth cameras simultaneously with rtabmap 3D SLAM

First, Download and Install rtabmap and rtabmap_ros following these instructions in the branch ros2.

In one terminal, launch the two cameras and rtabmap ros (make sure that you source the workspace where rtabmap_ros was built):

ros2 launch realsense_ros2 slam_rtabmap_launch.py

Published topics

rs_t265_node

• rs_t265/odom [nav_msgs/Odometry]: Pose and speeds of the t265 tracking camera.
• rs_t265/imu [sensor_msgs/Imu]: Imu data.
• rs_t265/fisheye1/image_raw [sensor_msgs/Image]: Left fisheye image (only published if publish_fisheye is set to true).
• rs_t265/fisheye2/image_raw [sensor_msgs/Image]: Right fisheye image (only published if publish_fisheye is set to true).
• rs_t265/fisheye1/camera_info [sensor_msgs/CameraInfo]: Calibration/intrinsics for left fisheye (only published if publish_fisheye is set to true).
• rs_t265/fisheye2/camera_info [sensor_msgs/CameraInfo]: Calibration/intrinsics for right fisheye (only published if publish_fisheye is set to true).
• tf

rs_d435_node

• rs_d435/point_cloud [sensor_msgs/PointCloud2]: Pointcloud from d435 depth camera.
• rs_d435/aligned_depth/image_raw [sensor_msgs/Image]: Depth Image from d435 depth camera (only published if publish_depth parameter is set to true ).
• rs_d435/image_raw [sensor_msgs/Image]: Raw Color images