# https://github.com/IntelRealSense/realsense-ros/tree/ros2-development
# first, ROS 2 does not work with conda environment, so we need to better have python 3.11
# ubuntu 24.04 gets you python 3.12
# laptop4 originally has python3.11 in conda and python3.12
# ROS2 says it cannot work with anaconda
https://docs.ros.org/en/jazzy/How-To-Guides/Using-Python-Packages.html
# you cannot change the system python version since apt-get relies on it
1. Install ROS2 Jazzy for Ubuntu 24 (conda deactivate, with system python3.12)
https://docs.ros.org/en/jazzy/Installation/Ubuntu-Install-Debs.html
# uninstall
sudo apt remove ~nros-jazzy-* && sudo apt autoremove
$ sudo apt install software-properties-common
$ sudo add-apt-repository universe
...
$ sudo apt install ros-jazzy-desktop
1.2. Setup Environment
# everytime you need to source the ros2 stuff
$ source /opt/ros/jazzy/setup.bash
1.3. Test
Screen 1: send some info as topic using C++ API
$ source /opt/ros/jazzy/setup.bash
$ ros2 run demo_nodes_cpp talker
Screen 2: receive the message using Python API
$ source /opt/ros/jazzy/setup.bash
$ ros2 run demo_nodes_py listener
# check your ROS2 is using the python version you normally use
$ ls /opt/ros/jazzy/lib/python3.12/
# the python package for ROS 2 is rclpy
# you will need to conda deactivate to import rclpy
>>> rclpy.__path__
['/opt/ros/jazzy/lib/python3.12/site-packages/rclpy']
2. Install Realsense SDK 2.0
# need to build from source on Ubuntu 24.04:
https://github.com/IntelRealSense/realsense-ros/tree/ros2-development?tab=readme-ov-file#option-3-build-from-source
https://github.com/IntelRealSense/librealsense/blob/master/doc/installation.md
# clone the github repo librealsense and install
# Note: The shared object will be installed in /usr/local/lib, header files in /usr/local/include.
# The binary demos, tutorials and test files will be copied into /usr/local/bin
# install - also build the python 3.12 wrapper
# install pyrealsense wrapper, for python3.12 we need to build from source\
# https://github.com/IntelRealSense/librealsense/tree/master/wrappers/python#building-from-source
$ mkdir build && cd build
$ cmake ../ -DCMAKE_BUILD_TYPE=Release -DBUILD_PYTHON_BINDINGS:bool=true -DPYTHON_EXECUTABLE=/usr/bin/python3
$ sudo make uninstall && make clean && make -j8 && sudo make install
# pyrealsense2 is installed to '/usr/local/lib/python3.12/dist-packages/pyrealsense2'
# copy librealsense/build/Release/*.so* lib to the system python, add __init__.py file
junweil@precognition-laptop4:~$ sudo cp ./projects/realsense/librealsense/build/Release/pyrealsense2.cpython-312-x86_64-linux-gnu.so* /usr/local/lib/python3.12/dist-packages/pyrealsense2/
junweil@precognition-laptop4:~/projects/realsense/librealsense/build/Release$ sudo vi /usr/local/lib/python3.12/dist-packages/pyrealsense2/__init__.py
# py libs (pyd/so) should be copied to pyrealsense2 folder
from .pyrealsense2 import *
$ ls /usr/local/lib/python3.12/dist-packages/pyrealsense2
__init__.py pyrsutils.cpython-312-x86_64-linux-gnu.so
pyrealsense2.cpython-312-x86_64-linux-gnu.so pyrsutils.cpython-312-x86_64-linux-gnu.so.2.55
pyrealsense2.cpython-312-x86_64-linux-gnu.so.2.55 pyrsutils.cpython-312-x86_64-linux-gnu.so.2.55.1
pyrealsense2.cpython-312-x86_64-linux-gnu.so.2.55.1
# if set -DPYTHON_EXECUTABLE=/home/junweil/anaconda3/bin/python
# pyrealsense2 is installed to /home/junweil/anaconda3/lib/python3.12/site-packages/pyrealsense2
# we need to install this for system python
# test:
realsense-viewer # check latency qualitatively
python - import pyrealsense2
3. Install Realsense ROS package
# install from source to have a latency test tool - nope, the latency is running locally, not useful
# install from apt install
# https://github.com/IntelRealSense/realsense-ros/tree/ros2-development?tab=readme-ov-file#option-1-install-debian-package-from-ros-servers-foxy-eol-distro-is-not-supported-by-this-option
$ sudo apt install ros-jazzy-realsense2-*
# this will install the realsense ros stuff to
# ['/opt/ros/jazzy']
#### note we may need to conda deactivate to use ROS python
Testing
# Realsense ROS 2 documentation: https://dev.intelrealsense.com/docs/ros2-wrapper
# code: https://github.com/IntelRealSense/realsense-ros
1. running the RGBD Topic in one process
# https://github.com/IntelRealSense/realsense-ros/tree/ros2-development?tab=readme-ov-file#rgbd-topic
$ conda deactivate
$ source /opt/ros/jazzy/setup.bash
junweil@precognition-laptop4:~/projects/realsense$ ros2 launch realsense2_camera rs_launch.py enable_rgbd:=true enable_sync:=true align_depth.enable:=true enable_color:=true enable_depth:=true depth_module.depth_profile:=640x480x30 rgb_camera.color_profile:=640x480x30 camera_namespace:=go2 camera_name:=d435i
# depth_qos:=SENSOR_DATA color_qos:=SENSOR_DATA
# so this can be best_effort, higher frame rate?
# depth_module.depth_profile:=640x480x30 rgb_camera.color_profile:=640x480x30
# camera_namespace:=robot1 camera_name:=D455_1
# so you get /robot1/D455_1 node
# to see this code: https://github.com/IntelRealSense/realsense-ros
# realsense2_camera/launch/rs_launch.py
$ ros2 topic list
/camera/camera/aligned_depth_to_color/camera_info
/camera/camera/aligned_depth_to_color/image_raw
/camera/camera/color/camera_info
/camera/camera/color/image_raw
/camera/camera/color/metadata
/camera/camera/depth/camera_info
/camera/camera/depth/image_rect_raw
/camera/camera/depth/metadata
/camera/camera/extrinsics/depth_to_color
/camera/camera/rgbd
/parameter_events
/rosout
/tf_static
# RGBD Message
std_msgs/Header header
sensor_msgs/CameraInfo rgb_camera_info
sensor_msgs/CameraInfo depth_camera_info
sensor_msgs/Image rgb
sensor_msgs/Image depth
# check the Hz
$ ros2 topic hz /camera/camera/aligned_depth_to_color/image_raw
2 Hz - 25 Hz??
$ ros2 topic hz /camera/camera/color/camera_info
30 Hz
$ ros2 topic hz /camera/camera/depth/image_rect_raw
22 Hz
$ ros2 topic hz /camera/camera/color/image_raw
2-12 Hz
# making it faster over ROS?
https://github.com/IntelRealSense/librealsense/issues/10730
2. listen to this message and visualize
# need to install a package (first do $ conda deactivate)
$ python3 -m pip install numpy-quaternion --break-system-packages
2.1 use the example code
junweil@precognition-laptop4:~/projects/realsense/realsense-ros/realsense2_camera/scripts$ python3 rs2_listener.py /camera/camera/color/image_raw
# https://github.com/IntelRealSense/realsense-ros/blob/ros2-development/realsense2_camera/scripts/rs2_listener.py
# show center depth
junweil@precognition-laptop4:~/projects/realsense/realsense-ros/realsense2_camera/scripts$ python3 show_center_depth.py
2.2 Run opencv to check the delay time,
# on the same machine, delay time is ~0.05 second
junweil@precognition-laptop4:~/projects/depth_cameras/depth_ros$ python3 image_listener.py
# video:
https://github.com/JunweiLiang/depth_ros/blob/main/ros2_realsense_depth_delay_test.avi
Realsense on Jetsons
Installation
1. Reboot the Jetson Nano 8GB on Go2 to set 25W power mode
# Ubuntu 20.04, Jetpack 5.1.1
# jtop to see status
# can connect from the laptop in the same LAN
$ ssh unitree@192.168.123.18
pwd: 123
2. Install ROS 2
# no need, go2 already has ROS 2
(go2) unitree@ubuntu:~$ ls /opt/ros/foxy/
# python 3.8
3. Install Realsense SDK
# use apt install, easy with network
https://github.com/IntelRealSense/librealsense/blob/master/doc/installation_jetson.md
3.1 install pyrealsense2
# 换清华源
pip config set global.index-url https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple
# (go2) unitree@ubuntu:~$ pip install pyrealsense2
3.2 测试pyrealsense under GUI
# clone this: https://github.com/JunweiLiang/tennis_project
(go2) unitree@ubuntu:~/projects/tennis_project$ python get_depth_images.py
# FPS 27+
4. Install Realsense ROS package
$ sudo apt install ros-foxy-realsense2-*
Testing
# Test with local ROS visualization and see the delay
1. run the data publisher in one screen session
(go2) unitree@ubuntu:~/projects/tennis_project$ ros2 launch realsense2_camera rs_launch.py enable_rgbd:=true enable_sync:=true align_depth.enable:=true enable_color:=true enable_depth:=true depth_module.depth_profile:=640x480x30 rgb_camera.color_profile:=640x480x30 camera_namespace:=go2 camera_name:=d435i
# depth_qos:=SENSOR_DATA color_qos:=SENSOR_DATA
# so this can be best_effort, we can get lower latency and higher framerate
# depth_module.depth_profile:=640x480x30 rgb_camera.color_profile:=640x480x30
# camera_namespace:=robot1 camera_name:=D455_1
# so you get /robot1/D455_1 node
# to see this code: https://github.com/IntelRealSense/realsense-ros
# realsense2_camera/launch/rs_launch.py
2. run the subscriber to visualize
# Jetson Nano换成25W模式，最大功率
# 注意狗子上的路由器，一定要接5V/3A的电源，5V/2A的电源，功率受限会经常断开wifi
# nano local run, ~0.05-0.06 second delay, 27 FPS
# cmd
(go2) unitree@ubuntu:~/projects/depth_ros$ python image_listener.py --save ros2_realsense_depth_delay_test_besteffort_ongo2nano.avi
# video
ros2_realsense_depth_delay_test_ongo2nano.avi
# on laptop4 with wired connection to LAN, we can directly see this ros topic (需要重启laptop4，才能有线连上路由器)
# 0.14 second delay, 29 FPS
# video: ros2_realsense_depth_delay_test_onlaptop4+overwiredgo2nano.avi
# best effort, got 0.0001 second delay??
# video: ros2_realsense_depth_delay_test_besteffor_onlaptop4+overwiredgo2nano.avi
# wifi test
# use a USB wifi connector, disable the laptop one
$ sudo ifconfig wlp59s0f0 down
# no need
# cmd
junweil@precognition-laptop4:~/projects/depth_cameras/depth_ros$ python3 image_listener.py --save ros2_realsense_depth_delay_test_besteffort_onlaptop4+overwifigo2nano.avi
# over wifi, FPS is ~10, , delay is 0.14 second, video:
ros2_realsense_depth_delay_test_onlaptop4+overwifigo2nano.avi
# best effort is FPS ~15, delay 0.11 second, video:
ros2_realsense_depth_delay_test_besteffort_onlaptop4+overwifigo2nano.avi
Orbbec on Ubuntu 24.04
Installation
# https://github.com/orbbec/OrbbecSDK_ROS2
# python sdk: https://github.com/orbbec/pyorbbecsdk
1. ROS 2 installation done
2. Install Orbbec ROS 2 package
# see full instruction here: https://github.com/orbbec/OrbbecSDK_ROS2?tab=readme-ov-file#installation-instructions
junweil@precognition-laptop4:~/projects/depth_cameras/orbbec_ros2_ws/src$ git clone https://github.com/orbbec/OrbbecSDK_ROS2.git
..
junweil@precognition-laptop4:~/projects/depth_cameras/orbbec_ros2_ws$ colcon build --event-handlers console_direct+ --cmake-args -DCMAKE_BUILD_TYPE=Release
3. Install Python SDK
# make a python wheel to install
https://github.com/orbbec/pyorbbecsdk?tab=readme-ov-file#making-a-python-wheel
# open3D does not support python3.12 yet, ignored
installed to
/home/junweil/.local/lib/python3.12/site-packages/pyorbbecsdk.cpython-312-x86_64-linux-gnu.so
Testing
# we use Orbbec Gemini 336L: https://www.orbbec.com/products/stereo-vision-camera/gemini-336l/
# ROS2 下掉帧？https://github.com/orbbec/OrbbecSDK_ROS2/blob/main/docs/fastdds_tuning.md
1. start the camera publishing
# need to source two thing
$ source ros2
junweil@precognition-laptop4:~/projects/depth_cameras/orbbec_ros2_ws$ . ./install/setup.bash
junweil@precognition-laptop4:~/projects/depth_cameras/orbbec_ros2_ws$ ros2 launch orbbec_camera gemini_330_series.launch.py
# other launch file
junweil@precognition-laptop4:~/projects/depth_cameras/orbbec_ros2_ws$ ls src/OrbbecSDK_ROS2/orbbec_camera/launch/
# we have these topics
junweil@precognition-laptop4:~/projects/depth_cameras/orbbec_ros2_ws$ ros2 topic list
/camera/color/camera_info
/camera/color/image_raw
/camera/color/metadata
/camera/depth/camera_info
/camera/depth/image_raw
/camera/depth/metadata
/camera/depth_filter_status
/camera/depth_to_color
/diagnostics
/parameter_events
/rosout
/tf
/tf_static
# launch params: https://github.com/orbbec/OrbbecSDK_ROS2?tab=readme-ov-file#launch-parameters
junweil@precognition-laptop4:~/projects/depth_cameras/orbbec_ros2_ws$ ros2 launch orbbec_camera gemini_330_series.launch.py color_width:=640 color_height:=480 color_fps:=30 depth_width:=640 depth_height:=480 depth_fps:=30 enable_color:=true enable_depth:=true depth_registration:=true enable_spatial_filter:=false enable_temporal_filter:=false
# 640x480x30 profile and depth align to color:
color_width:=640 color_height:=480 color_fps:=30 depth_width:=640 depth_height:=480 depth_fps:=30 enable_color:=true enable_depth:=true depth_registration:=true
# best effort [10/09/2024] this does not work, we need to run the default profile
color_qos:=SENSOR_DATA depth_qos:=SENSOR_DATA
# enable_spatial_filter:=true enable_temporal_filter:=true does not seem to help
# compressed image?
https://github.com/orbbec/OrbbecSDK_ROS2?tab=readme-ov-file#compressed-image
# getting synchronized image and depth
https://github.com/orbbec/OrbbecSDK_ROS2/issues/37#issuecomment-2149734469
2. run visualization
junweil@precognition-laptop4:~/projects/depth_cameras/depth_ros$ python3 image_listener_orbbec.py
