QVideo: Qt framework for video cameras

QVideo is a framework for integrating video cameras into Qt projects (PyQt6 or PyQt5; PySide6 is experimental) for scientific research. It provides a unified, registration-based property system so that every camera backend — USB webcams, GenICam devices, FLIR cameras, Raspberry Pi cameras — is controlled through the same API. Property trees, display widgets, and a digital video recorder are built on top of that abstraction and require no camera-specific code.

Features

Unified camera API — QCamera subclasses expose adjustable parameters via registerProperty / registerMethod; UI and recording layers consume them without knowing the underlying hardware.
Auto-built property trees — QCameraTree reads the registered property map and builds a pyqtgraph parameter tree widget automatically.
Threaded video source — QVideoSource wraps any camera in a QThread and emits newFrame(ndarray, FrameInfo) at acquisition rate.
Composable filter pipeline — VideoFilter objects run in a QFilterStack on a worker thread between source and consumers (display, recorder); filters include blur, edge detection, RGB channel selection, sample-and-hold, binary threshold, and blob coloring.
Graphical overlays — QTrackpyWidget for live particle tracking and QYoloWidget for real-time object detection render markers directly on QVideoScreen; composite mode lets the DVR record the annotated scene.
Digital video recorder — lossless HDF5 (with timestamps) and OpenCV video formats; QDVRWidget is the composite UI widget.
Live display — QVideoScreen supports mouse-aware graphical overlays for annotations, regions of interest, and user interaction.

Installation

pip install "QVideo[pyqt6]"

Qt binding

QVideo needs a Qt binding: PyQt6 (recommended) or PyQt5. The binding is an optional extra so that QVideo can use one you already have, such as the system packages on a Raspberry Pi, instead of installing a second copy:

pip install "QVideo[pyqt6]"   # PyQt6 (recommended)
pip install "QVideo[pyqt5]"   # PyQt5
pip install QVideo            # use a binding that is already installed

PySide6 is experimental (pip install "QVideo[pyside6]"): it runs in continuous integration, but not all of QVideo's tests have been adapted to it.

Raspberry Pi

PyPI has no prebuilt ARM wheels for PyQt5 or PyQt6, so the system-provided packages must be used instead. Create the virtual environment with --system-site-packages so pip can see them:

sudo apt install python3-pyqt6 python3-pyqt6.qtsvg
python3 -m venv --system-site-packages .venv
source .venv/bin/activate
pip install QVideo

Replace python3-pyqt6 with python3-pyqt5 if you prefer PyQt5.

Optional hardware backends

Backend	Extra	Notes
GenICam cameras (Vimba, etc.)	`pip install QVideo[genicam]`	Requires a vendor-supplied `.cti` producer file
Raspberry Pi camera	`pip install QVideo[picamera]`	Requires `picamera2`

Quick start

import pyqtgraph as pg
from QVideo.cameras.Noise import QNoiseSource
from QVideo.lib import QVideoScreen

pg.mkQApp()

source = QNoiseSource()          # synthetic noise — no hardware needed
screen = QVideoScreen()
source.newFrame.connect(screen.setImage)

screen.show()
source.start()
pg.exec()

Replace QNoiseSource with QOpenCVSource, QGenicamSource, etc. to switch hardware — the rest of the code is identical.

Camera backends

Backend	Class	Hardware
`cameras/Noise`	`QNoiseCamera`	Synthetic — no hardware required
`cameras/OpenCV`	`QOpenCVCamera`	USB webcams via OpenCV
`cameras/Genicam`	`QGenicamCamera`	Abstract base for all GenICam/GigE Vision cameras
`cameras/Flir`	`QFlirCamera`	FLIR cameras via GenICam (Spinnaker GenTL producer)
`cameras/Basler`	`QBaslerCamera`	Basler cameras via GenICam (pylon GenTL producer)
`cameras/IDS`	`QIDSCamera`	IDS Imaging cameras via GenICam
`cameras/MV`	`QMVCamera`	Any GenICam camera via MATRIX VISION mvGenTLProducer
`cameras/Vimbax`	`QVimbaXCamera`	Allied Vision cameras via VimbaX GenTL producer
`cameras/Picamera`	`QPicamera`	Raspberry Pi camera module

Compatibility

Verified on real hardware (the unit tests mock all hardware):

Backend	macOS	Ubuntu	Raspberry Pi OS	Windows
OpenCV	14 – 26	24.04	tested	not tested
FLIR (GenICam)	14 – 26	24.04	—	not tested

On macOS 26.7.1 (Tahoe), the OpenCV backend was tested with OpenCV 5.0.0_11 and the FLIR backend with Spinnaker SDK 4.4.0.246. Other combinations are expected to work but have not been checked. See the documentation for details.

QVideo itself is pure Python and is tested on Python 3.11 – 3.14. The newest Python release may lag: Python 3.15 is run in CI as an allowed-failure job because some dependencies (e.g. scipy) do not yet publish wheels for it.

Writing a new camera backend

Subclass QCamera and implement three methods:

from QVideo.lib import QCamera

class MyCamera(QCamera):

    def _initialize(self) -> bool:
        self.device = open_my_hardware()
        if not self.device:
            return False
        self.registerProperty('exposure',
                              getter=lambda: self.device.get_exposure(),
                              setter=lambda v: self.device.set_exposure(v),
                              ptype=float)
        return True

    def _deinitialize(self) -> None:
        self.device.close()

    def read(self):
        ok, frame = self.device.read_frame()
        return ok, frame

QCameraTree and QVideoSource work with MyCamera immediately — no additional code needed.

Filters

Filter	Class	Description
Gaussian blur	`QBlurFilter`	Smoothing with adjustable kernel radius
Canny edge detection	`QEdgeFilter`	Edge map with configurable thresholds
RGB channel selection	`QRGBFilter`	Pass one or more color channels
Sample and hold	`QSampleHold`	Background normalization via a sampled median estimate
Binary threshold	`QThresholdFilter`	Convert to binary mask at a configurable level
Blob coloring	`QBlobFilter`	Color connected foreground regions with distinct hues
YOLO annotation	`QYOLOFilter`	Annotate frames with YOLO bounding boxes (requires `ultralytics`)

Overlays

Overlays render analysis results directly on the live QVideoScreen. Each needs its own optional packages: pip install QVideo[trackpy] for the trackpy overlay, pip install QVideo[yolo] for the YOLO overlay (which pulls in PyTorch), or pip install QVideo[overlays] for both.

Overlay	Class	Description
Particle tracking	`QTrackpyWidget`	Live particle detection and tracking using `trackpy`
Object detection	`QYoloWidget`	Real-time bounding-box detection using YOLO (`ultralytics`)

Acknowledgements

Work on this project at New York University is supported by the National Science Foundation of the United States under award number DMR-2428983 and by an award from the TAC Program of New York University.

Metadata

Release files for QVideo 4.3.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for QVideo 4.3.0
File	Size	Uploaded	Provenance
qvideo-4.3.0.tar.gz	277.6 kB	Oct 25, 2026	With provenance	Details

Built distribution (wheel)

Table of built distributions (wheels) for QVideo 4.3.0
File	Interpreter	ABI	Platform	Reset
qvideo-4.3.0-py3-none-any.whl 217.2 kB Oct 25, 2026 With provenance	Python 3	none	any	Details

Total release size: 494.7 kB

Release files / qvideo-4.3.0.tar.gz

Download URL	qvideo-4.3.0.tar.gz
Size	277.6 kB
Tags	Source
SHA-256 checksum How to use checksums	`ee0cb184c8ee86e32fa74d916537791c15c5b9a9a0a7059cc075606813872ee2`
BLAKE2b-256 checksum How to use checksums	`b5aad63f158c20f670e8cab16f8f54b4cf22b296d2a147bc0004204b08983b85`
Upload date	Oct 25, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	Yes
Uploaded via	`twine/7.0.0 CPython/3.13.16`

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Identity

Publishing platform: GitHub Actions
Publishing repository: github.com/davidgrier/QVideo
Publishing commit: github.com/davidgrier/QVideo/tree/4c30f603c33e9b7ccda5afd86923d3c663fa9577

Workflow

Publishing configuration: github.com/davidgrier/QVideo/blob/4c30f603c33e9b7ccda5afd86923d3c663fa9577/.github/workflows/publish.yml
Publishing logs: github.com/davidgrier/QVideo/actions/runs/40219428693/attempts/1

Artifact

Subject: qvideo-4.3.0.tar.gz
SHA-256 checksum: ee0cb184c8ee86e32fa74d916537791c15c5b9a9a0a7059cc075606813872ee2 Verifying attestations

Release files / qvideo-4.3.0-py3-none-any.whl

Download URL	qvideo-4.3.0-py3-none-any.whl
Size	217.2 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`2a8783f4ba15e0bc2d47a177f902fa25009f7442a831a1e8c291c7f1f5195827`
BLAKE2b-256 checksum How to use checksums	`051942d17ef8617c63863477405cf8ca33bd199326faeac85dc93d4b555cd3f3`
Upload date	Oct 25, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	Yes
Uploaded via	`twine/7.0.0 CPython/3.13.16`

Provenance

Identity

Publishing platform: GitHub Actions
Publishing repository: github.com/davidgrier/QVideo
Publishing commit: github.com/davidgrier/QVideo/tree/4c30f603c33e9b7ccda5afd86923d3c663fa9577

Workflow

Publishing configuration: github.com/davidgrier/QVideo/blob/4c30f603c33e9b7ccda5afd86923d3c663fa9577/.github/workflows/publish.yml
Publishing logs: github.com/davidgrier/QVideo/actions/runs/40219428693/attempts/1

Artifact

Subject: qvideo-4.3.0-py3-none-any.whl
SHA-256 checksum: 2a8783f4ba15e0bc2d47a177f902fa25009f7442a831a1e8c291c7f1f5195827 Verifying attestations

Security for QVideo 4.3.0

Report project as malware

Provenance

Full provenance

All 2 files in this release have been uploaded with provenance metadata.

Attestation for qvideo-4.3.0.tar.gz

Identity

Publishing platform: GitHub Actions
Publishing repository: github.com/davidgrier/QVideo
Publishing commit: github.com/davidgrier/QVideo/tree/4c30f603c33e9b7ccda5afd86923d3c663fa9577

Workflow

Publishing configuration: github.com/davidgrier/QVideo/blob/4c30f603c33e9b7ccda5afd86923d3c663fa9577/.github/workflows/publish.yml
Publishing logs: github.com/davidgrier/QVideo/actions/runs/40219428693/attempts/1

Artifact

Subject: qvideo-4.3.0.tar.gz
SHA-256 checksum: ee0cb184c8ee86e32fa74d916537791c15c5b9a9a0a7059cc075606813872ee2 Verifying attestations

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.