How to build ZED 2i Camera x ROS2 Foxy x Nvidia Jetson x Ubuntu 18.04 via Docker

This article documents a Docker-based approach for integrating the ZED 2i stereo camera with ROS2 Foxy on Nvidia Jetson hardware running Ubuntu 18.04.

Key Motivation

The author explains two critical drivers for this work:

• ROS2 Eloquent reached end-of-life; Foxy represents the stable long-term support alternative
• “Officially, ZED 2i has no launch file at Eloquent branch”

Core Challenges Identified

The implementation proved difficult due to several factors:

• Limited available documentation
• ROS2 Foxy targets Ubuntu 20.04, not 18.04
• Missing shared libraries in NVIDIA’s L4T base image
• Conflicting information across online resources

The Dockerfile Solution

The author provides a complete Docker configuration that:

• Bases on codustry/ros:foxy-ros-base-l4t-r32.5.0
• Downloads and installs ZED SDK 3.5 for Jetpack 4.5
• Configures environment variables for CUDA and ROS
• Applies L4T binaries and patches
• Builds the ZED ROS2 wrapper and diagnostics packages

Key configuration includes symbolic linking and specific CMake arguments to resolve linking issues.

Failed Approaches

The author documents three unsuccessful methods:

1. OS upgrade from Ubuntu 18.04 to 20.04 (corrupts package manager)
2. Building L4T from base images (complexity with CUDA integration)
3. Starting with Ubuntu 20.04 (incompatible with Jetson drivers)

Docker Advantages Highlighted

• QEMU enables parallel, local development
• Cloud builds leverage faster internet connectivity
• Layer-by-layer inspection capability
• Faster deployment than SD card flashing
• Complete reproducibility

Final Result

The working setup launches successfully with:

ros2 launch zed_wrapper zed2i.launch.py