DEPZ Stereo Vision Logo

Documentation du SDK des capteurs

Documentation des SDK de capteurs USB DEPZ en Python, JavaScript (TypeScript), C#, Java, C, C++ et Rust : installation, guides pratiques avec exemples, références complètes de l'API — des pages dédiées pour chaque capteur, le SDK Android et le paquet ROS 2.

Voir sur GitHub

depz_sensors is the ROS 2 package for the DEPZ USB sensor line: one node that finds every DEPZ board on USB and publishes its measurements as standard sensor_msgs messages. It runs on ROS 2 Jazzy (Ubuntu 24.04) and ROS 2 Lyrical (Ubuntu 26.04) and is built on the C++ SDK, which does the USB and sensor work — you write no C++ to use it.

ros2 launch depz_sensors sensors.launch.py

Every board gets topics of its own, named after its model and USB serial number, so they stay the same across restarts and replugs:

/depz/hcsr04_uexample01/range      sensor_msgs/msg/Range
/depz/vl53l8ch_texample01/depth    sensor_msgs/msg/Image
/depz/bno086_iexample01/imu        sensor_msgs/msg/Imu
/depz/bno086_iexample01/mag        sensor_msgs/msg/MagneticField

What each board publishes

Board Topic Message Default rate
HC-SR04 range sensor_msgs/msg/Range, metres 20 Hz
VL53L4CD · VL53L0X · VL53L1CX · VL53L1CB · VL53L3CX · VL53L4CX range sensor_msgs/msg/Range, metres about 20 Hz
VL53L5CX · VL53L7CX · VL53L7CH · VL53L8CX · VL53L8CH depth sensor_msgs/msg/Image, 32FC1, metres per zone 15 Hz, 8x8
TMF8829 depth sensor_msgs/msg/Image, 32FC1, metres per zone 30 Hz, 8x8
BNO055 · BNO085 · BNO086 imu, mag sensor_msgs/msg/Imu, sensor_msgs/msg/MagneticField 100 Hz

Units follow REP 103 (metres, radians, rad/s, m/s², tesla). A measurement the sensor could not make is +inf (nothing in range) or NaN (rejected by the sensor, or a zone without a target) — REP 117 — never a made-up distance. Messages are published with the sensor-data QoS (best effort) and stamped with the moment of the measurement: the board's clock, mapped to the computer's.

Quick start

  1. Install ROS 2 (Jazzy on Ubuntu 24.04) and build the package from the SDK repository with colcon (Install has every step, ROS 2 Lyrical and Docker):

    source /opt/ros/jazzy/setup.bash
    git clone https://github.com/depz-ai/depz-sensor-sdk.git ~/depz-sensor-sdk
    mkdir -p ~/depz_ws && cd ~/depz_ws
    colcon build --base-paths ~/depz-sensor-sdk/packages/depz-sensor-ros2
    source install/setup.bash
  2. Let your user open the boards — install the udev rule the package ships (it also keeps ModemManager off the ports), then replug the boards:

    sudo cp $(ros2 pkg prefix --share depz_sensors)/udev/60-depz-sensors.rules /etc/udev/rules.d/
    sudo udevadm control --reload-rules && sudo udevadm trigger
  3. Start the node and look at the data from a second terminal:

    ros2 launch depz_sensors sensors.launch.py
    ros2 topic list
    ros2 topic echo /depz/hcsr04_uexample01/range

    Or see every board at once in RViz 2, the 3D viewer of ROS — a cone per distance sensor, a box that turns with each IMU, each multizone sensor's zone grid — with one command that starts the node and RViz together:

    ros2 launch depz_sensors view.launch.py

No ROS on the machine? The package's scripts/ros.sh jazzy run does the same in a Docker container (Install).

Names of your own

Give boards aliases by serial number in a parameter file, and pass it with config:=:

/**:
  ros__parameters:
    aliases:
      - "UEXAMPLE01=front"
      - "IEXAMPLE01=imu"
    hot_plug: true
ros2 launch depz_sensors sensors.launch.py config:=$HOME/my_sensors.yaml

The topics are then /depz/front/range and /depz/imu/imu, the frames front and imu; hot_plug picks up boards plugged in later.

From Python and C++

The package installs example programs to start from — distance, every distance sensor in one table, depth grid, the closest zone (in Python and the same in C++), IMU angles:

ros2 run depz_sensors range_listener.py --ros-args -p topic:=/depz/front/range
ros2 run depz_sensors range_table.py
ros2 run depz_sensors depth_listener.py --ros-args -p topic:=/depz/vl53l8ch_texample01/depth
ros2 run depz_sensors depth_min_distance.py --ros-args -p topic:=/depz/vl53l8ch_texample01/depth
ros2 run depz_sensors depth_min_distance --ros-args -p topic:=/depz/vl53l8ch_texample01/depth
ros2 run depz_sensors imu_listener.py --ros-args -p topic:=/depz/imu/imu

example_robot.launch.py starts the node and examples together and says where each board is mounted on the robot (a static transform from base_link to the board's frame) — the shape of a launch file for your robot. Examples has each program and what it prints.

Subscribe with qos_profile_sensor_data (C++: rclcpp::SensorDataQoS()): a subscriber that asks for reliable delivery gets nothing from a best-effort publisher.

Where to next

  • Install — step by step: ROS 2 Jazzy or Lyrical, the build with colcon, Docker instead of ROS on the machine, USB permissions.
  • Guide — running the node, viewing depth in rqt_image_view and every board in RViz 2 (view.launch.py), what each message carries, aliases, every parameter, coordinate frames and static transforms, time stamps, troubleshooting by the line the node prints, limitations.
  • Examples — the programs above one by one, with what each prints, and a launch file that starts them with the node.
  • C++ SDK — the library under the node, for programs that talk to a board directly.
  • Source & license — open source (MIT): github.com/depz-ai/depz-sensor-sdk; the package's README is these pages in one.

Nous respectons votre vie privée

Nous utilisons des cookies pour améliorer votre expérience de navigation, diffuser du contenu personnalisé et analyser notre trafic. En cliquant sur « Tout accepter », vous consentez à notre utilisation des cookies. Voir la politique de cookies