(EMGU) 如何拆分和合并图像？

Question

I am working in C# on Visual Studio with Emgu.

I am doing a several image manipulations on a large image. I had the idea of splitting the image in half, doing the manipulations in parallel, them merging the image.

In pursuit of this goal, I have found a number of questions regarding the acquisition of rectangular parts of images for processing as well as splitting an image into channels (RGB, HSV, etc). I have not found a question that addresses the task of taking an image, and making it into two images. I have also not found a question that addresses taking two images and tacking them together.

The following code is what I would like to do, where split and merge are imaginary methods to accomplish it.

Image<Bgr,Byte> ogImage = new Image<Bgr, byte>(request.image);
Image<Bgr,Byte> topHalf = new Image<Bgr, byte>();
Image<Bgr,Byte> bottomHalf = new Image<Bgr, byte>();

ogImage.splitHorizonally(topHalf,bottomHalf);

//operations

ogImage = topHalf.merge(bottomHalf);

This is the type of question I hate asking, because it is simple and you would think it has a simple, easily available solution, but I have not found it, or I have found it and not understood it.

Answer 1

There are a number of ways to solve this but here is what I did. I took the easiest way out;-)

    Mat lena = new Mat(@"D:\OpenCV\opencv-3.2.0\samples\data\Lena.jpg", 
                       ImreadModes.Unchanged);

    CvInvoke.Imshow("Lena", lena);

    System.Drawing.Rectangle topRect = new Rectangle(0, 
                                                     0, 
                                                     lena.Width,
                                                     (lena.Height / 2));

    System.Drawing.Rectangle bottomRect = new Rectangle(0, 
                                                        (lena.Height / 2),
                                                        lena.Width,
                                                        (lena.Height / 2));

    Mat lenaTop = new Mat(lena, topRect);

    CvInvoke.Imshow("Lena Top", lenaTop);

    Mat lenaBottom = new Mat(lena, bottomRect);

    CvInvoke.Imshow("Lena Bottom", lenaBottom);

    Mat newLena = new Mat();

    CvInvoke.VConcat(lenaBottom, lenaTop, newLena);

    CvInvoke.Imshow("New Lena", newLena);

    CvInvoke.WaitKey(0);

Original Lena

Lena Top Half

Lena Bottom Half

The New Lena Rearranged

Answer 2

Your goal isn't splitting an image. Your goal is to parallelize some operation on the image.

You did not disclose the specific operations you need to perform. That is important to know however, if you want to parallelize those operations.

You need to learn about strategies for parallelization in general. Commonly, a "kernel" is executed on several partitions of the data in parallel.

One practical approach is called OpenMP . You apply "pragmas" to your own loops and OpenMP spreads those loop iterations across different threads.