simplepycam-tspspi 0.1.7

Simple Python native access library for Video4Linux camera devices

Simple Python native access library for Video4Linux camera devices

Warning: Work (slowly) in progress. Currently hardwired to YUV format. This will be fixed soon

This is a simple library that allows one to access camera devices accessible via Video4Linux from Python without using libraries with huge dependency chains that allows easy and fast testing of ideas and algorithms for image processing before implementing them in a proper way in some programming language such as C. It’s built in the most simple way on purpose.

Tested on:

• Python 3.8:

  • FreeBSD 12 (amd64, aarch64)

  • FreeBSD 13 (amd64, aarch64)

Currently not implemented / not optimal

• Currently the library does not support anything else than YUV422 captures from the cameras and only outputs RGB888 images as nested Python lists. This will be fixed as soon as possible but will introduce a breaking change.

Installing via PyPi

Note that you have to install some native packages on FreeBSD before building the extension. Those are:

• multimedia/v4l_compat

• multimedia/webcamd

To install and start the required services:

pkg install multimedia/v4l_compat multimedia/webcamd
echo "webcamd_enable=\"YES\"" >> /etc/rc.conf
echo "cuse_load=\"YES\"" >> /boot/loader.conf
kldload cuse
/etc/rc.d/devd restart
/usr/local/etc/rc.d/webcamd start

The package is available as a source distribution via PyPi:

pip install simplepycam-tspspi

Building

Use setup.py in cext:

$ python setup.py build

Example usage

More sophisticated examples can be found in the samples directory.

Using the stream callback interface, the default format and with

import simplepycam

def processFrame(camera, frame):
    if shouldStopProcessing:
        return False
    else:
        return True

with simplepycam.Camera("/dev/video0") as cam:
    cam.frameCallback = [ processFrame ]
    cam.stream()

Using the stream callback interface, the default format and open/close

import simplepycam

def processFrame(camera, frame):
    if shouldStopProcessing:
        return False
    else:
        return True

cam = simplepycam.Camera("/dev/video0")
cam.open()
cam.frameCallback = [ processFrame ]
cam.stream()
cam.close()

Using the polling API, the default format and with

import simplepycam

with simplepycam.Camera("/dev/video0") as cam:
    cam.streamOn()
    for i in range(100):
        frame = cam.nextFrame()
    cam.streamOff()

Using the polling API, the default format and open/close

import simplepycam

cam = simplepycam.Camera("/dev/video0")
cam.open()
cam.streamOn()
for i in range(100):
    frame = cam.nextFrame()
cam.streamOff()
cam.close()

API documentation

UML diagram of Python side