ZED Foxglove Integration

A multi-camera streaming application that captures data from Stereolabs ZED cameras and streams it to Foxglove for real-time visualization and recording. The project includes both Python and C++ implementations for maximum flexibility.

Prerequisites

Installation

Python Version

1. Install ZED SDK following the official installation guide

2. Navigate to the Python directory:

   cd python/

3. Install Python dependencies using Poetry:

   # Install Poetry if you haven't already
   curl -sSL https://install.python-poetry.org | python3 -
   
   # Install dependencies
   poetry install

   Or using pip:

   pip install foxglove-sdk>=0.4.0 numpy>=2.0 opencv-python>=4.10 requests>=2.31

4. Install ZED Python API: The project includes pyzed-5.0-cp310-cp310-linux_x86_64.whl. Install it with:

   pip install pyzed-5.0-cp310-cp310-linux_x86_64.whl

C++ Version

1. Install ZED SDK as above

2. Navigate to the C++ directory:

   cd cpp/

3. Build the project:

   mkdir build && cd build
   cmake ..
   make -j$(nproc)

Usage

Python Version

cd python/

# Stream to Foxglove Studio via WebSocket
poetry run python main.py --ws

# Record to MCAP file
poetry run python main.py --mcap output.mcap

# Both streaming and recording
poetry run python main.py --ws --mcap output.mcap

C++ Version

cd cpp/build/

# Stream to Foxglove Studio via WebSocket
./zed_foxglove --ws

# Record to MCAP file
./zed_foxglove --mcap output.mcap

# Both streaming and recording
./zed_foxglove --ws --mcap output.mcap

Command Line Options

Option	Description	Default
--ws	Enable WebSocket server for live streaming	Disabled
--mcap <file>	Output MCAP file name	output.mcap (C++), disabled (Python)
--help, -h	Show help message	-

Foxglove Studio Configuration

1. Start Foxglove Studio

2. Connect to live stream:

   • Click "Open connection"
   • Select "Foxglove WebSocket"
   • Use URL: ws://localhost:8765

3. Open recorded data:

   • Click "Open local file"
   • Select your .mcap file

4. Add panels to visualize:

   • Image Panel: Subscribe to /image_0, /image_1, etc. for camera images
   • Image Panel: Subscribe to /depth_0, /depth_1, etc. for depth maps
   • 3D Panel: Subscribe to /point_cloud_0, /point_cloud_1, etc. for point clouds