rpi4_rs

Wireless networking for Intel Realsense Cameras connected to Raspberry Pi 4 (Model B) units.

TODO List

• frames-per-second optimization (replaced pickle with zlib)
• include instrinsic data with frames
• parsing realsense config files

Hardware Requirements

• Intel Realsense camera (tested on D455)
• Raspberry Pi 4 (Model B)
• Wireless Router (tested on a TP-Link Archer C80)

RPi4 Setup

This repository assumes your Raspberry Pi 4 unit uses Ubuntu 22.04 instead of the original Raspbian OS.

1. Ensure you have the tools needed to build the librealsense SDK on your Pi 4 unit.

sudo apt install python3 python3-pip
sudo apt install cmake build-essential

2. Run the librealsense SDK installation script.

chmod +x ./libuvc_installation.sh
./libuvc_installation.sh

3. The Pi 4 unit will be running the Server.py script, which requires NumPy.

pip3 install numpy

Client Setup

This repository assumes your client is running on a Linux machine with AMD64.

1. Install the librealsense SDK following the guide from the librealsense repository here. The dkms and utils packages are sufficient enough.

2. Install Python prerequisites, and the Realsense Python Wrapper.

pip3 install numpy opencv-python pyrealsense2

Network Setup

Connect your Pi 4 unit to your wireless router, ensuring it uses a static ip address. Here's a simple guide to setting a static ip address for your device.

In client.py and server.py, change the mc_ip_address variable to the static ip assigned to your Pi 4 unit.

Running the Client and Server

1. Streaming RGB and Depth

1. On your Pi 4, run the Server.py script.
2. On your client device, run the Client.py script.

Run the script in the exact sequence outlined above.

2. Streaming RGB only

1. On your Pi 4, run the ServerRGB.py script.
2. On your client device, run the ClientRGB.py script.

Acknowledgement

The client-server scripts are adapted from EtherSense and the librealsense SDK installation script is adapted from librealsense. Thanks to krehov100 for the open-source ethernet networking for Intel Realsense cameras.