cuda(0)

　　自从进入电气工程，阴差阳错的学习了cuda，在学习的过程中初步了解并行计算的一些思路，

由于我的电脑空间有限，所以我准备记录平时的一些cuda编程的练习，放到博客园上这样方便我

以后的学习，这样也可以为博客园的成长尽一下绵薄之力。接下来进入正题:

　　图像灰度转换cuda实现：

#include "cuda_runtime.h"
#include "device_launch_parameters.h"
#include <stdio.h>
#include<iostream>

#include <iostream>  
#include <opencv2/core/core.hpp>  
#include <opencv2/highgui/highgui.hpp> 

#define CHANNELS 3

__global__ void colorToGreyScaleConversion(unsigned char *pout, unsigned char *pin, int width, int height) {
    int Col = blockIdx.x*blockDim.x + threadIdx.x;
    int Row = blockIdx.y*blockDim.y + threadIdx.y;

    if (Col < width && Row < height) {
        int greyoffset = Row * width + Col;
        int rgbOffset = greyoffset * CHANNELS;

        unsigned char r = pin[rgbOffset];
        unsigned char g = pin[rgbOffset + 1];
        unsigned char b = pin[rgbOffset + 2];

        pout[greyoffset] = 0.21f*r + 0.71f*g + 0.07f*b;
    }
}

using namespace cv;
int main(void) {

    // 读入一张图片（缩小图）    
    Mat img = imread("E:\\opencv\\lena512color.tiff");
    // 创建一个名为 "图片"窗口    
    namedWindow("lena");
    // 在窗口中显示图片   
    imshow("lena", img);
    // 等待6000 ms后窗口自动关闭    
    waitKey(6000);

    const int imgheight = img.rows;
    const int imgwidth = img.cols;
    const int imgchannel = img.channels();

    Mat grayImage(imgheight, imgwidth, CV_8UC1, Scalar(0));

    unsigned char *dev_pin;
    unsigned char *dev_pout;

    cudaMalloc((void**)&dev_pin, imgheight*imgwidth*imgchannel* sizeof(unsigned char));
    cudaMalloc((void**)&dev_pout, imgheight*imgwidth*sizeof(unsigned char));

    cudaMemcpy(dev_pin, img.data, imgheight*imgwidth*imgchannel * sizeof(unsigned char), cudaMemcpyHostToDevice);


    dim3 BlockDim(16, 16);
    dim3 GridDim((imgwidth - 1) / BlockDim.x + 1, (imgheight - 1) / BlockDim.y + 1);
    colorToGreyScaleConversion << <GridDim, BlockDim >> > (dev_pout, dev_pin, imgwidth, imgheight);

    cudaMemcpy(grayImage.data, dev_pout, imgheight*imgwidth*sizeof(unsigned char), cudaMemcpyDeviceToHost);

    cudaFree(dev_pin);
    cudaFree(dev_pout);
    imshow("grayImage", grayImage);
    waitKey(3000);
    return 0;
}

　　值得注意的是：我是在vs2017平台上实现的，在编程的时候调用了openCV的库函数，读者需要自行配置该环境，为了不让

大家在配置环境时耗费太长时间，我附上CSDN上一位大神的配置教程链接：https://blog.csdn.net/qq_41175905/article/details/80560429

转换前如下图：

　　

转换后如下图：

另外读者也可用Matlab, openCV等库函数实现，比较发现，cuda并行计算的时间要快一些。