rpi-vision

Capture snapshots from a webcam connected to a Raspberry Pi and send them to a Mac on the same local network.

The Raspberry Pi acts as the camera device, while the Mac receives the snapshots for viewing, storage, or later processing.

USB webcam → Raspberry Pi 5 → local network → Mac

Project status

The Python prototype captures a frame, keeps a temporary lossless PNG on the Pi, and can send an in-memory JPEG asynchronously to a receiver on the Mac.

See ARCHITECTURE.md for the current format and data-flow decisions.

See docs/SAME_WIFI_WORKFLOW.md for the complete Mac receiver, firewall, network-checking, and continuous-transfer procedure.

For an ordered review of the whole repository, start with review/README.md.

Goals

• Capture still images from the USB webcam connected to the Raspberry Pi.
• Send snapshots to a Mac on the same Wi-Fi or Ethernet network.
• Keep the Raspberry Pi setup lightweight and compatible with ARM64.
• Provide a simple, reproducible Python environment for development.

Hardware

The current test machine is a Raspberry Pi 5 Model B with:

• a four-core ARM Cortex-A76 CPU;
• 4 GB of RAM;
• a 64 GB-class microSD card;
• a generic USB 2.0 camera; and
• 64-bit Debian GNU/Linux.

No NVIDIA GPU or CUDA runtime is required for webcam capture and snapshot transfer.

See HARDWARE.md for the full hardware, operating-system, camera, network, storage, and compatibility inventory.

Quick start

The Python implementation is a reusable local package managed with uv:

cd python-stream
uv sync --locked
uv run python -m unittest discover -s tests -v
uv run rpi-vision capture --help

See python-stream/README.md for its public API, commands, local-path installation, and package layout.

Repository layout

rpi-vision/
├── README.md
├── HARDWARE.md
├── ARCHITECTURE.md
├── PLAN.md
├── review/
├── python-stream/
├── cpp-stream/
└── lessons/

Generated virtual environments, bytecode, tool caches, and build artifacts are excluded from version control.