NVIDIA Jetson Nano with Intel RealSense Depth Camera Using ROS2 Humble

Autonomous Machines Jetson & Embedded Systems Jetson Nano
1

NVIDIA Jetson Nano with Intel RealSense Depth Camera Using ROS2 Humble

1280×720 68.9 KB

In this tutorial, we’ll explore how to interface an NVIDIA Jetson Nano with an Intel RealSense Depth Camera using ROS2 Humble. This setup is powerful for robotics applications that require real-time perception and processing capabilities. The Jetson Nano provides the computational power to handle the data from the RealSense camera, while ROS2 Humble offers a robust framework for developing and managing robotics software.

1500×1129 195 KB

Prerequisites

Before we start, ensure you have the following components:

  • NVIDIA Jetson Nano with Ubuntu 20.04 OS image
  • Intel RealSense Depth Camera (e.g., D435i)
  • ROS2 Humble installed on the Jetson Nano
  • USB 3.0 cable for connecting the RealSense camera to the Jetson Nano
  • Internet connection for downloading necessary packages

Installing ROS2 RealSense Package

sudo apt install ros-humble-realsense2-camera

Launch the RealSense Node:

Create a launch file to start the RealSense node. Create a new file realsense_launch.py:

from launch import LaunchDescription
from launch_ros.actions import Node

def generate_launch_description():
return LaunchDescription([
Node(
package=‘realsense2_camera’,
executable=‘realsense2_camera_node’,
name=‘realsense2_camera’,
output=‘screen’,
parameters=[{
‘enable_depth’: True,
‘enable_infra1’: True,
‘enable_infra2’: True,
‘enable_color’: True,
}],
),
])

Run the Launch File:

1500×844 473 KB

ros2 launch your_package_name realsense_launch.py

Visualizing the Data in rqt:

1500×844 130 KB

Add RealSense Data to rqt:

1500×844 119 KB

  • Add a new DepthCloud display and set the topic to /camera/depth/color/points.
  • Add an Image display and set the topic to /camera/color/image_raw

1500×844 126 KB

Depth image

What is a Depth Image?

A depth image is a grayscale image where each pixel represents the distance from the camera to the objects in the scene. Unlike a regular RGB image, which contains color information, a depth image encodes depth information, allowing you to perceive the 3D structure of the environment.

Key Topics and Messages

In ROS 2, the depth image from a RealSense camera is published on a specific topic, commonly /camera/depth/image_raw. The message type for the depth image is sensor_msgs/Image. Here’s a breakdown of the main elements:

  • Topic: /camera/depth/image_raw
  • Message Type: sensor_msgs/Image

sensor_msgs/Image Message

The sensor_msgs/Image message contains several fields, but the most relevant for depth images are:

  • header: Standard ROS message header containing a timestamp and coordinate frame information.
  • height: Height of the image in pixels.
  • width: Width of the image in pixels.
  • encoding: Encoding type of the image data, e.g., 16UC1 for a 16-bit unsigned single-channel image.
  • is_bigendian: Whether the image data is stored in big-endian byte order.
  • step: Full row length in bytes.
  • data: Actual pixel data, stored as a byte array.

Processing Depth Images

Depth images can be processed in various ways, depending on your application. Common tasks include:

  • Object Detection and Recognition: Identifying and classifying objects based on their 3D shape.
  • Obstacle Avoidance: Using depth information to detect and avoid obstacles in the robot’s path.
  • 3D Mapping: Creating a 3D map of the environment for navigation purposes.

Conclusion

By following these steps, you should now have an NVIDIA Jetson Nano interfaced with an Intel RealSense Depth Camera using ROS2 Humble. This setup is highly useful for various robotics applications, enabling you to leverage real-time depth and color data for navigation, object detection, and more. Experiment with different ROS2 nodes and packages to expand the capabilities of your robotic system.

Feel free to comment below if you encounter any issues or have further questions!

1 Like