Raspberry Pi 5 + Camera Module 3 Setup for ROS 2 Jazzy (Ubuntu 24.04)

This repository provides a comprehensive guide and necessary commands to set up the official Raspberry Pi Camera Module 3 on Raspberry Pi 5 running Ubuntu 24.04 LTS and ROS 2 Jazzy.

The Issue: The standard ROS 2 installation (sudo apt install ros-jazzy-camera-ros) installs the upstream version of libcamera, which currently lacks full support for the RPi 5 + Camera Module 3 combination on Ubuntu. This results in "no cameras available" errors.

The Solution: We need to manually build the Raspberry Pi fork of libcamera from source and link it with the camera_ros node within a colcon workspace.

Sources:

• Based on the official Raspberry Pi libcamera fork: https://github.com/raspberrypi/libcamera
• Camera ROS node wrapper: https://github.com/christianrauch/camera_ros

---

1. System Preparation (On Raspberry Pi)

Update the system and install all necessary build tools, dependencies, and the ROS 2 colcon-meson plugin.

sudo apt install -y g++ cmake meson ninja-build pkg-config 
sudo apt install -y libboost-dev libglib2.0-dev libgstreamer-plugins-base1.0-dev libgnutls28-dev openssl 
sudo apt install -y libjpeg-dev libpng-dev libtiff-dev 
sudo apt install -y python3-yaml python3-ply pybind11-dev 
sudo apt install -y qt6-base-dev libqt6widgets6 
sudo apt install -y python3-colcon-meson 

2. Create ROS2 workspace (if you already haven't).

Create standard folder structure for ros2_ws.

mkdir -p ~/ros2_ws/src

3. Clone source code.

Clone the specific Raspberry Pi fork of libcamera and the ROS 2 camera node.

cd ~/ros2_ws/src

# 1. Clone the RPi fork of libcamera
git clone https://github.com/raspberrypi/libcamera.git

# 2. Clone the ROS 2 camera node wrapper
git clone https://github.com/christianrauch/camera_ros.git

4. Building.

Install any remaining ROS dependencies and build the packages. Note that camera_ros will automatically link against our local libcamera build.

While being in ~/ros2_ws:

# Check for missing dependencies
rosdep install --from-paths src --ignore-src -y

# Build the packages (this may take 5-15 minutes on RPi 5)
colcon build --packages-select libcamera camera_ros

Hint: Make sure that your hardware setup is correct, CSI camera port is enabled. (It should be by default).

5. Running the Camera Node (On Raspberry PI).

After rebooting, connect via SSH and run the node. Always source the local setup file to use the custom-built library.

cd ~/ros2_ws
source install/setup.bash
ros2 run camera_ros camera_node

Warning: After running node you will probably get load of INFO and WARNING logs. Don't mind them as long as there are is no FATAL/ERROR. The easiest test is to just test it using ROS2's topic list command.

Hint: You may encounter ERROR corresponding to calibration files, but since it's your first test and you haven't created them it isn't a problem.

ros2 topic list

# Expected result: 
/camera/camera_info
/camera/image_raw
/camera/image_raw/compressed
/parameter_events
/rosout

6. Viewing image.

To view the video stream on your laptop (connected to the same Wi-Fi):

1. Make sure you have installed rqt_image_view.

sudo apt install ros-jazzy-rqt-image-view ros-jazzy-image-transport-plugins

2. Run the viewer.

source /opt/ros/jazzy/setup.bash
ros2 run rqt_image_view rqt_image_view

3. Select topic from dropdown menu (left top corner):

Select /camera/image_raw/compressed.

7. Common Issues & Troubleshooting (Viewing image).

rqt_image_view Crashing or Freezing

Symptoms: The application closes unexpectedly with invalid allocator error or freezes when refreshing the topic list. Cause: This is a known bug in the rqt_image_view application in ROS 2 when handling compressed streams during topic switching. Workaround:

1. Restart the application.
2. If it keeps crashing, try selecting the empty row in the topic list (or a different topic) before switching to /camera/image_raw/compressed.

Terminal Errors: (!buf.empty())

Symptoms: You see repeated red error logs in the terminal like: OpenCV(4.6.0) ... error: (-215:Assertion failed) !buf.empty() in function 'imdecode_' Cause: This is normal when streaming video over Wi-Fi. It indicates that a data packet was lost during transmission (Packet Loss). Solution: Ignore it. As long as the video stream in the window looks smooth, the system is working correctly. ROS 2 skips the broken frame and displays the next one.