Skip to content

Depth Camera ROS 2 Package Usage (Aurora933)

Hint

This document provides a standalone description of the build, launch, topic inspection, RViz display, and parameter configuration of the Aurora930 / Aurora933 depth camera ROS 2 package.

  • Robot side: Ubuntu 22.04 + ROS 2 Humble
  • Depth camera: Deptrum / Guangjian Aurora930 series; the car uses Aurora933
  • ROS 2 driver source package: deptrum-ros-driver-aurora930-aarch64-0.2.11-source.tar.gz
  • Build option: -DSTREAM_SDK_TYPE=AURORA930
  • Default namespace: /aurora
  • Main outputs: RGB images, IR images, Depth images, and PointCloud2 point clouds
  • The main difference between Aurora930 and Aurora933 is that the former has no glass dust cover, while the latter has a dust-proof glass cover.

Package/resource archive download link: (refer to Kit resources/Resource links to download)

Scenario Description

This package connects Aurora930 series depth cameras to a ROS 2 system and publishes RGB, IR, Depth, point cloud, and camera intrinsic topics for subsequent functions such as object detection, depth localization, robotic arm grasping, obstacle perception, and RViz visualization.

After the ROS 2 driver is started, the common outputs are as follows:

  • /aurora/rgb/image_raw: Color image, suitable for object detection, QR code recognition, and image preview.
  • /aurora/depth/image_raw: Depth image, encoded as 16-bit depth data and usually treated as millimeter-level depth.
  • /aurora/ir/image_raw: Infrared image, used for infrared imaging and debugging.
  • /aurora/points2: Point cloud data, used for 3D display and spatial localization in RViz.
  • /aurora/rgb/camera_info, /aurora/ir/camera_info: Camera intrinsic information.
Attention

Before starting the ROS 2 driver, it is recommended to complete the USB enumeration, udev permission, and SDK example test steps described in the hardware test document. If the SDK example cannot open the camera, the ROS 2 driver usually cannot open the camera normally either.

The ROS 2 package-related content in the archive is as follows:

930_SDK等文件/
└── 930_ROS/
    └── ROS2/
        └── ros2源码/
            └── deptrum-ros-driver-aurora930-aarch64-0.2.11-source.tar.gz

The main structure of the extracted ROS 2 source package is as follows:

deptrum-ros-driver-aurora930-0.2.11/
├── CMakeLists.txt
├── package.xml
├── docs/
│   ├── building.md
│   └── usage_aurora930.md
├── launch_aurora930/
│   └── launch/
│       ├── aurora930_launch.py
│       ├── aurora930_multi_launch.py
│       ├── viewer930_launch.py
│       └── sub_node_ci_aurora930_launch.py
├── rviz/
│   └── aurora930-ros2.rviz
├── src/
├── include/
├── subscribe_node/
└── ext/

The commonly used files are described below:

File / Directory Purpose
aurora930_launch.py Starts the Aurora930 depth camera driver node
viewer930_launch.py Starts RViz and loads the official RViz configuration
aurora930_multi_launch.py Multi-device launch reference
sub_node_ci_aurora930_launch.py Starts the example node that subscribes to and saves images
rviz/aurora930-ros2.rviz RViz display configuration for Aurora930
subscribe_node/ Example program for image subscription and saving

Step 1: Install Basic ROS 2 Dependencies

In an Ubuntu 22.04 + ROS 2 Humble environment, it is recommended to install the common dependencies first:

sudo apt update
sudo apt install -y \
  build-essential \
  cmake \
  python3-colcon-common-extensions \
  libusb-1.0-0-dev \
  libudev-dev \
  libopencv-dev \
  libgoogle-glog-dev \
  libgflags-dev \
  ros-humble-cv-bridge \
  ros-humble-image-transport \
  ros-humble-camera-info-manager \
  ros-humble-tf2 \
  ros-humble-tf2-ros \
  ros-humble-tf2-geometry-msgs \
  ros-humble-angles \
  ros-humble-rviz2 \
  ros-humble-rqt-image-view

Source the ROS 2 environment:

source /opt/ros/humble/setup.bash
Attention

If the system is not Humble, replace humble in the commands with the name of the current ROS 2 distribution.

Step 2: Prepare the ROS 2 Workspace

Create a workspace:

mkdir -p ~/dev_ws/src
cd ~/dev_ws/src

Copy the ROS 2 source archive from the resource package to ~/dev_ws/src and extract it:

tar -xzf deptrum-ros-driver-aurora930-aarch64-0.2.11-source.tar.gz

After extraction, the directory is usually:

deptrum-ros-driver-aurora930-0.2.11

You can rename it to a shorter directory name:

mv deptrum-ros-driver-aurora930-0.2.11 deptrum-ros-driver

The recommended final structure is as follows:

~/dev_ws/
└── src/
    └── deptrum-ros-driver/
        ├── CMakeLists.txt
        ├── package.xml
        ├── launch_aurora930/
        ├── rviz/
        ├── src/
        ├── include/
        ├── subscribe_node/
        └── ext/

Step 3: Confirm the ROS 2 Package Name

The package name used by the launch files in the driver source is:

deptrum-ros-driver-aurora930

If the package name in package.xml is not deptrum-ros-driver-aurora930, it must be changed.

Enter the source directory:

cd ~/dev_ws/src/deptrum-ros-driver

Check the package name:

grep "<name>" package.xml

If the output is not the following:

<name>deptrum-ros-driver-aurora930</name>

You can perform a one-time replacement:

sed -i 's/deptrum-ros-driver/deptrum-ros-driver-aurora930/g' package.xml
Attention

Run the sed replacement command only once. Before running it, confirm the package name with grep "<name>" package.xml to avoid an invalid package name caused by repeated replacement.

Step 4: Build the ROS 2 Package

Return to the workspace root directory:

cd ~/dev_ws
source /opt/ros/humble/setup.bash

Build for the Aurora930 type:

colcon build --cmake-args -DSTREAM_SDK_TYPE=AURORA930

If you have switched the SDK type or modified the CMake configuration, you can force a reconfigure:

colcon build --cmake-force-configure --cmake-args -DSTREAM_SDK_TYPE=AURORA930

After the build is complete, source the workspace environment:

source ~/dev_ws/install/setup.bash

Check whether the package is visible:

ros2 pkg list | grep deptrum

Normally, the following should be displayed:

deptrum-ros-driver-aurora930

Step 5: Start the Depth Camera Driver

Run the following in a terminal:

source /opt/ros/humble/setup.bash
source ~/dev_ws/install/setup.bash
ros2 launch deptrum-ros-driver-aurora930 aurora930_launch.py

The namespace is set by default in the driver launch file:

namespace="aurora"

Therefore, the actual topic names are prefixed with /aurora.

After a successful launch, open another terminal to inspect the topics:

source /opt/ros/humble/setup.bash
source ~/dev_ws/install/setup.bash
ros2 topic list | grep aurora

A common output is as follows:

/aurora/rgb/image_raw
/aurora/rgb/camera_info
/aurora/ir/image_raw
/aurora/ir/camera_info
/aurora/depth/image_raw
/aurora/points2

Check the frame rate of the depth image:

ros2 topic hz /aurora/depth/image_raw

Check the point cloud topic information:

ros2 topic info /aurora/points2

Step 6: RViz Visualization

Option 1: Start with the Official RViz Configuration

The driver package already contains the RViz configuration file for Aurora930, so you can start it directly:

source /opt/ros/humble/setup.bash
source ~/dev_ws/install/setup.bash
ros2 launch deptrum-ros-driver-aurora930 viewer930_launch.py

This command opens RViz and loads:

aurora930-ros2.rviz

Option 2: Open RViz Manually

If you need to configure RViz manually, run:

ros2 run rviz2 rviz2

Configure RViz as follows:

  1. Set Fixed Frame under Global Options to:
depth_camera_link
  1. Click Add, select By topic, and add the following display items:
/aurora/depth/image_raw
/aurora/ir/image_raw
/aurora/rgb/image_raw
/aurora/points2
Attention

If you run RViz directly in an SSH terminal, qt.qpa.xcb: could not connect to display may appear. This is a graphical display environment issue, not a camera driver issue. It is recommended to run RViz on a PC on the same ROS 2 network, or use a remote desktop / X11 forwarding.

Step 7: View Images with rqt_image_view

If you only want to view images and do not need the point cloud, you can use rqt_image_view:

ros2 run rqt_image_view rqt_image_view

After starting it, select the topic to view:

/aurora/rgb/image_raw
/aurora/depth/image_raw
/aurora/ir/image_raw

Where:

  • The RGB image is used to confirm whether the color camera image is normal.
  • The Depth image is used to confirm whether the depth data is normal.
  • The IR image is used to confirm whether the infrared module is normal.

Step 8: Save Image Data

The driver source contains the subscribe_node example program, which can subscribe to the topics published by the driver and save RGB, IR, and Depth images locally.

After starting the depth camera driver, run the following in another terminal:

source /opt/ros/humble/setup.bash
source ~/dev_ws/install/setup.bash
cd ~/dev_ws/install/deptrum-ros-driver-aurora930/lib/deptrum-ros-driver-aurora930
./sub_node

The default save path is:

/tmp/image

You can also use launch to start the subscriber node:

ros2 launch deptrum-ros-driver-aurora930 sub_node_ci_aurora930_launch.py
Attention

The depth camera driver node must be started before saving images; otherwise, the subscriber node has no data source.

Common Launch Parameters

aurora930_launch.py provides several adjustable parameters that can be overridden on the command line at launch time.

Parameter Name Default Value Purpose
rgb_enable true Whether to publish RGB color images
ir_enable true Whether to publish IR infrared images
depth_enable true Whether to publish depth images
rgbd_enable false Whether to enable RGBD mode
point_cloud_enable true Whether to publish point cloud data; depends on the depth image
boot_order 1 Camera startup order configuration
ir_fps 15 IR frame rate; selectable values are usually 5 / 10 / 12 / 15
rgb_fps 15 RGB frame rate; should be less than or equal to the IR frame rate
exposure_enable false Whether to set exposure manually
exposure_time 1 Exposure time; the range is usually 1 to 31
gain_enable false Whether to set gain manually
gain_value 20 Gain value; the range is usually 10 to 160
usb_port_number Empty Specifies the USB port number to distinguish multiple devices
threshold_size 110 Denoising filter size
depth_correction true Whether to enable depth correction
align_mode true Whether to enable RGBD / IR alignment mode
laser_power 1.0 Laser power mode; common meanings: 1 auto, 2 indoor, 3 outdoor
minimum_filter_depth_value 150 Minimum depth filter value
maximum_filter_depth_value 4000 Maximum depth filter value
resolution_mode_index 2 Resolution mode index
log_dir /tmp/ Log output directory
stream_sdk_log_enable true Whether to enable SDK logs
heart_enable false Whether to enable heartbeat detection

Start only the RGB and depth images and do not publish point clouds:

ros2 launch deptrum-ros-driver-aurora930 aurora930_launch.py \
  rgb_enable:=true \
  ir_enable:=false \
  depth_enable:=true \
  point_cloud_enable:=false

Limit the maximum valid depth to 3 meters:

ros2 launch deptrum-ros-driver-aurora930 aurora930_launch.py \
  maximum_filter_depth_value:=3000

Reduce the load and keep only the depth image:

ros2 launch deptrum-ros-driver-aurora930 aurora930_launch.py \
  rgb_enable:=false \
  ir_enable:=false \
  depth_enable:=true \
  point_cloud_enable:=false

ROS 2 Topic Description

Because namespace="aurora" is set in the launch file, the complete topic names are as follows:

Full Topic Name Message Type Purpose
/aurora/rgb/image_raw sensor_msgs/msg/Image RGB color image, BGR888 / bgr8 format
/aurora/rgb/camera_info sensor_msgs/msg/CameraInfo RGB camera intrinsics
/aurora/ir/image_raw sensor_msgs/msg/Image IR infrared image
/aurora/ir/camera_info sensor_msgs/msg/CameraInfo IR / Depth camera intrinsics
/aurora/depth/image_raw sensor_msgs/msg/Image Depth image, 16-bit depth data
/aurora/points2 sensor_msgs/msg/PointCloud2 Depth point cloud

If other nodes subscribe to /camera/depth/image_raw or /camera/color/image_raw by default, you can connect them by remapping:

ros2 run your_package your_node --ros-args \
  -r /camera/color/image_raw:=/aurora/rgb/image_raw \
  -r /camera/depth/image_raw:=/aurora/depth/image_raw \
  -r /camera/color/camera_info:=/aurora/rgb/camera_info \
  -r /camera/depth/camera_info:=/aurora/ir/camera_info

Multi-Device Usage

If multiple ROS 2 devices are on the same LAN or the same host, you can set different ROS_DOMAIN_ID values to avoid confusion in DDS communication.

For example:

export ROS_DOMAIN_ID=5

If you need the setting to take effect permanently:

echo "export ROS_DOMAIN_ID=5" >> ~/.bashrc
source ~/.bashrc

For multiple depth cameras, you can also specify the device port with usb_port_number so that the driver does not open the wrong device.

Log Viewing

By default, the driver outputs logs to the /tmp/ directory. You can view the latest log symlink:

ls -l /tmp/node.INFO

View the log contents:

tail -f /tmp/node.INFO

You can also locate the driver log file:

ls /tmp | grep deptrum_ros_driver

For complete usage, it is recommended to open three terminals.

First Terminal: Start the Depth Camera Driver

source /opt/ros/humble/setup.bash
source ~/dev_ws/install/setup.bash
ros2 launch deptrum-ros-driver-aurora930 aurora930_launch.py

Second Terminal: View Topics and Frame Rate

source /opt/ros/humble/setup.bash
source ~/dev_ws/install/setup.bash
ros2 topic list | grep aurora
ros2 topic hz /aurora/depth/image_raw

Third Terminal: Image or Point Cloud Visualization

View images:

ros2 run rqt_image_view rqt_image_view

View the point cloud:

ros2 launch deptrum-ros-driver-aurora930 viewer930_launch.py

Image 1