While runing 3 camera's using this taking almost 50% of CPU

[rajeshamudala]

in our app we are using this library and running 3 camera's , in that 2 camera's are monochrome camera's and one is RGB camera.
when they are running amost 50% of CPU is occupying

• Each camera is running on different thread

• modified your usbcamera.cs class which will return byte[] after some processing steps, They are

• Fliiping the camera feed , Compressing the image quality and converting to byte[]

• Here is the modified class

       ' public byte[] BufferToBitmap_Bytes_V2(byte[] buffer)
        {
            if (buffer == null || buffer.Length == 0)
                return null;
  
            try
            {
                // Step 1: Create OpenCV Mat directly from the buffer
                using (var mat = new OpenCvSharp.Mat(Height, Width, OpenCvSharp.MatType.CV_8UC3, buffer))
                {
                    // Step 2: Flip the Mat vertically
                    //OpenCvSharp.Cv2.Flip(mat, mat, OpenCvSharp.FlipMode.X);
  
                    // Step 3: Compress the Mat to JPEG
                    int[] compressionParams =
                    {
            (int)OpenCvSharp.ImwriteFlags.JpegQuality,
            OCT_Config.IRIS_Image_Quality_For_Transfer
        };
  
                    OpenCvSharp.Cv2.ImEncode(".jpg", mat, out var jpegData, compressionParams);
  
                    return jpegData;
                }
            }
            catch (Exception ex)
            {
                // Log the error for debugging purposes
                ConsoleWrapper.WriteLine($"Error in BufferToBitmap_Bytes: {ex.Message}");
                return null;
            }
        }`
  

Can we optimize the performance and is there anyway to run the calculations on GPU

[secile]

as to using GPU, please see issue #44.
as to supports monochrome camera, please see issue #45.

[secile]

As I check your modified BufferToBitmap_Bytes_V2().

• original BuffetToBitmap is designed to return non compressed image data, but you compress image to jpeg file.
  (this is your modified version, it is okey.)

• compressing image may require some CPU power. You had better do not compress image?

• You are now using OpenCvSharp function.
  I think it may because you have to read monochrome image type(called Y8 or Y800).
  I have never use monochrome Y800 image format, but as I googled, it is not so difficult to read data.
  If you do not use OpenCvSharp and read Y800 by yourself, it may be possible to reduce CPU usage.

• As I told issue How to set the video feed and still capture to a single channel (monochrome) #45, I have a plan to support Y800 future,
  but currently, I do not have monochrome USB camera.
  After I get monochrome USB camera, I will try.