qamlib 0.8.0

A library for interacting with Video4Linux2 (V4L2) video devices

qamlib

This is a library meant to be an easy way to interact with a V4L2 camera, by having a simple interface to capture images, and change camera controls. It is a C++20 class (+ a few structs) together with Python bindings via pybind11.

While the Python bindings are the main usage of qamlib, it is also possible to build qamlib as a (static) C++ library.

Supported features

• Get/set "normal" and extended V4L2 controls.
• Get/set image formats.
• Get/set framesize and cropping.
• Get/set framerate (FPS).
• Read out frames from a capture video device as a NumPy array. With or without buffering.
• Subscribe to events for a V4L2 device.

Planned features

• Pushing frames to a output video device.
• Supporting multiplane devices.

There are also some features supported that are currently exclusive to Qtechnology A/S cameras, but these are not compiled when main-line kernel headers are detected.

Example

import qamlib

cam = qamlib.Camera("/dev/video0")

# Use context manager to start and stop streaming
with cam:
    metadata, frame = cam.get_frame() # gets an image as raw bytes
    # process image

See more in the documentation

Building

Python

Building the Python module is done via mesonpy.

Dependencies

• gcc
• libstdc++-dev
• meson
• mesonpy
• nlohmann-json
• pybind11
• pybind11_json
• python3-build
• python3-dev

To build the module:

python -m build

C++

Dependicies

• gcc
• meson
• ninja
• nlohmann-json

To build the library we start by running meson setup:

meson setup build -Dpython=false

Then to compile do

meson compile -C build

To install the package.

meson install

Testing

Under tests/ are some tests, these have only been actually tested on Qtechnology A/S cameras.