V4L2 (Video for Linux version 2) implemented by .NET Core. The required minimum version of .NET Core is 2.1. Theoretically support all Linux devices running .NET Core (.NET Core JIT depends on ARMv7 instructions).

dotnet/iot#664

1. Install v4l-utils.

   sudo apt-get install v4l-utils
   

2. Install System.Drawing native dependencies.

   sudo apt-get install libc6-dev libgdiplus libx11-dev
   

1. Create a VideoConnectionSettings and set the parameters for capture.

   VideoConnectionSettings settings = new VideoConnectionSettings(busId: 0)
   {
       CaptureSize = (2560, 1920),
       PixelFormat = PixelFormat.JPEG,
       ExposureType = ExposureType.Auto
   };

2. Create a communications channel to a video device.

   using VideoDevice device = VideoDevice.Create(settings);

3. Capture static image

   // Capture static image
   device.Capture("/home/pi/jpg_direct_output.jpg");
   
   // Change capture setting
   device.Settings.PixelFormat = PixelFormat.YUV420;
   
   // Get image stream, convert pixel format and save to file
   MemoryStream ms = device.Capture();
   Color[] colors = VideoDevice.Yv12ToRgb(ms, settings.CaptureSize);
   Bitmap bitmap = VideoDevice.RgbToBitmap(settings.CaptureSize, colors);
   bitmap.Save("/home/pi/yuyv_to_jpg.jpg", System.Drawing.Imaging.ImageFormat.Jpeg);

4. Other methods

   // Get the supported formats of the device
   IEnumerable<PixelFormat> formats = device.GetSupportedPixelFormats();
   // Get the resolutions of the format
   IEnumerable<(uint Width, uint Height)> resolutions = device.GetPixelFormatResolutions(PixelFormat.YUYV));
   // Query v4l2 controls default and current value
   VideoDeviceValue value = device.GetVideoDeviceValue(VideoDeviceValueType.Rotate);

cd V4l2.Samples
dotnet publish -c release -r linux-arm -o YOUR_FOLDER
sudo dotnet YOUR_FOLDER/V4l2.Samples.dll

Before build docker image, you need to modify SDK, runtime and apt sources(in China) to adapt to the corresponding Linux platform.

docker build -t v4l2-sample -f Dockerfile .
docker run --rm -it --device /dev/video0 -v /home/pi/images:/home/pi/images v4l2-sample