线程池给你写好了，想加速拿来用就行哈

 Mat pano;
 Ptr<Stitcher> stitcher = Stitcher::create(mode);
 Stitcher::Status status = stitcher->stitch(imgs, pano)

#include "t.h"
#include <unistd.h>
#include <stdlib.h>
#include <stdio.h>
#include <vector>
#include <opencv2/opencv.hpp>
#include <iostream>
#include <opencv2/highgui/highgui.hpp>  
#include"opencv2/imgproc/imgproc.hpp"
#include "opencv2/imgcodecs.hpp"
#include "opencv2/stitching.hpp"
#include "omp.h"
using namespace cv;
using namespace std;

Mat img=imread("./1.png");

//线程池初始化
void threadpool_init(threadpool_t *pool, int threads);

//往线程池中加入任务
void threadpool_add_task(threadpool_t *pool, void *(*run)(void *arg), void *arg);

//摧毁线程池
void threadpool_destroy(threadpool_t *pool);

bool divide_images = false;
Stitcher::Mode mode = Stitcher::PANORAMA;
vector<Mat> imgs;
string result_name = "result.jpg";

void printUsage(char** argv);
int parseCmdArgs(int argc, char** argv);

//任意添加自己的代码实现
void* mytask(void *arg)
{
    clock_t start,end;
    start=clock();

    Mat pano;
    Ptr<Stitcher> stitcher = Stitcher::create(mode);
    Stitcher::Status status = stitcher->stitch(imgs, pano);

    if (status != Stitcher::OK)
    {
        cout << "Can't stitch images, error code = " << int(status) << endl;
        exit(0);
    }

    imwrite(result_name, pano);
    cout << "stitching completed successfully\n" << result_name << " saved!";

    free(arg);
    end=clock();
    cout<<"图像拼接时间: "<<(double)(end-start)/CLOCKS_PER_SEC<<" s"<<endl;
    return NULL;
}
//测试代码
int main(int argc, char* argv[])
{   
    clock_t start,end;
    start=clock();

    int retval = parseCmdArgs(argc, argv);
    if (retval) return EXIT_FAILURE;

    threadpool_t pool;
    //初始化线程池，最多三个线程
    threadpool_init(&pool, 10);
   // int i;
    //创建十个任务
   //for(i=0; i < 10; i++)
  // {

        int *arg = new int((sizeof(int)));
        *arg = 0;
        threadpool_add_task(&pool, mytask, arg);


  //  }


    threadpool_destroy(&pool);
        end=clock();
    cout<<"多线程运行时间: "<<(double)(end-start)/CLOCKS_PER_SEC<<" s"<<endl;

    return EXIT_SUCCESS;
    //return 0;
}


void printUsage(char** argv)
{
    cout <<
         "Images stitcher.\n\n" << "Usage :\n" << argv[0] <<" [Flags] img1 img2 [...imgN]\n\n"
         "Flags:\n"
         "  --d3\n"
         "      internally creates three chunks of each image to increase stitching success\n"
         "  --mode (panorama|scans)\n"
         "      Determines configuration of stitcher. The default is 'panorama',\n"
         "      mode suitable for creating photo panoramas. Option 'scans' is suitable\n"
         "      for stitching materials under affine transformation, such as scans.\n"
         "  --output <result_img>\n"
         "      The default is 'result.jpg'.\n\n"
         "Example usage :\n" << argv[0] << " --d3 --try_use_gpu yes --mode scans img1.jpg img2.jpg\n";
}


int parseCmdArgs(int argc, char** argv)
{
     clock_t start,end;
    start=clock();

    if (argc == 1)
    {
        printUsage(argv);
        return EXIT_FAILURE;
    }

    for (int i = 1; i < argc; ++i)
    {   
        //查看帮助
        if (string(argv[i]) == "--help" || string(argv[i]) == "/?")
        {
            printUsage(argv);
            return EXIT_FAILURE;
        }
        //在像素较大时候，开启这个模式
        else if (string(argv[i]) == "--d3")
        {
            divide_images = true;
        }
        else if (string(argv[i]) == "--output")
        {
            result_name = argv[i + 1];
            i++;
        }
        else if (string(argv[i]) == "--mode")
        {  
            //仅仅是重叠度高的可用
            if (string(argv[i + 1]) == "panorama")
                mode = Stitcher::PANORAMA;
            //实际测试 scans 模式比 panorama 适用范围更为广泛
            else if (string(argv[i + 1]) == "scans")
                mode = Stitcher::SCANS;
            else
            {
                cout << "Bad --mode flag value\n";
                return EXIT_FAILURE;
            }
            i++;
        }
        else
        {   


            //终端读取一系列图片
            Mat img = imread(argv[i]);
            if (img.empty())
            {
                cout << "Can't read image '" << argv[i] << "'\n";
                return EXIT_FAILURE;
            }
            //对图片进行裁剪
            if (divide_images)
            {
                Rect rect(0, 0, img.cols / 2, img.rows);
                imgs.push_back(img(rect).clone());
                rect.x = img.cols / 3;
                imgs.push_back(img(rect).clone());
                rect.x = img.cols / 2;
                imgs.push_back(img(rect).clone());
            }
            else
                imgs.push_back(img);


        }
    }
    end=clock();
    cout<<"图像读取时间: "<<(double)(end-start)/CLOCKS_PER_SEC<<" s"<<endl;
    return EXIT_SUCCESS;
}

#ifndef _CONDITION_H_
#define _CONDITION_H_

#include <pthread.h>

//封装一个互斥量和条件变量作为状态
typedef struct condition
{
    pthread_mutex_t pmutex;
    pthread_cond_t pcond;
}condition_t;

//对状态的操作函数
int condition_init(condition_t *cond);
int condition_lock(condition_t *cond);
int condition_unlock(condition_t *cond);
int condition_wait(condition_t *cond);
int condition_timedwait(condition_t *cond, const struct timespec *abstime);
int condition_signal(condition_t* cond);
int condition_broadcast(condition_t *cond);
int condition_destroy(condition_t *cond);

#endif

#include "c.h"

//初始化
int condition_init(condition_t *cond)
{
    int status;
    if((status = pthread_mutex_init(&cond->pmutex, NULL)))
        return status;

    if((status = pthread_cond_init(&cond->pcond, NULL)))
        return status;

    return 0;
}

//加锁
int condition_lock(condition_t *cond)
{
    return pthread_mutex_lock(&cond->pmutex);
}

//解锁
int condition_unlock(condition_t *cond)
{
    return pthread_mutex_unlock(&cond->pmutex);
}

//等待
int condition_wait(condition_t *cond)
{
    return pthread_cond_wait(&cond->pcond, &cond->pmutex);
}

//固定时间等待
int condition_timedwait(condition_t *cond, const struct timespec *abstime)
{
    return pthread_cond_timedwait(&cond->pcond, &cond->pmutex, abstime);
}

//唤醒一个睡眠线程
int condition_signal(condition_t* cond)
{
    return pthread_cond_signal(&cond->pcond);
}

//唤醒所有睡眠线程
int condition_broadcast(condition_t *cond)
{
    return pthread_cond_broadcast(&cond->pcond);
}

//释放
int condition_destroy(condition_t *cond)
{
    int status;
    if((status = pthread_mutex_destroy(&cond->pmutex)))
        return status;

    if((status = pthread_cond_destroy(&cond->pcond)))
        return status;

    return 0;
}

#ifndef _THREAD_POOL_H_
#define _THREAD_POOL_H_

//线程池头文件

#include "c.h"

//封装线程池中的对象需要执行的任务对象
typedef struct task
{
    void *(*run)(void *args);  //函数指针，需要执行的任务
    void *arg;              //参数
    struct task *next;      //任务队列中下一个任务
}task_t;


//下面是线程池结构体
typedef struct threadpool
{
    condition_t ready;    //状态量
    task_t *first;       //任务队列中第一个任务
    task_t *last;        //任务队列中最后一个任务
    int counter;         //线程池中已有线程数
    int idle;            //线程池中kongxi线程数
    int max_threads;     //线程池最大线程数
    int quit;            //是否退出标志
}threadpool_t;


//线程池初始化
void threadpool_init(threadpool_t *pool, int threads);

//往线程池中加入任务
void threadpool_add_task(threadpool_t *pool, void *(*run)(void *arg), void *arg);

//摧毁线程池
void threadpool_destroy(threadpool_t *pool);

#endif

#include "t.h"
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <errno.h>
#include <time.h>

//创建的线程执行
void *thread_routine(void *arg)
{
    struct timespec abstime;
    int timeout;
    printf("thread %d is starting\n", (int)pthread_self());
    threadpool_t *pool = (threadpool_t *)arg;
    while(1)
    {
        timeout = 0;
        //访问线程池之前需要加锁
        condition_lock(&pool->ready);
        //空闲
        pool->idle++;
        //等待队列有任务到来 或者 收到线程池销毁通知
        while(pool->first == NULL && !pool->quit)
        {
            //否则线程阻塞等待
            printf("thread %d is waiting\n", (int)pthread_self());
            //获取从当前时间，并加上等待时间， 设置进程的超时睡眠时间
            clock_gettime(CLOCK_REALTIME, &abstime);  
            abstime.tv_sec += 2;
            int status;
            status = condition_timedwait(&pool->ready, &abstime);  //该函数会解锁，允许其他线程访问，当被唤醒时，加锁
            if(status == ETIMEDOUT)
            {
                printf("thread %d wait timed out\n", (int)pthread_self());
                timeout = 1;
                break;
            }
        }

        pool->idle--;
        if(pool->first != NULL)
        {
            //取出等待队列最前的任务，移除任务，并执行任务
            task_t *t = pool->first;
            pool->first = t->next;
            //由于任务执行需要消耗时间，先解锁让其他线程访问线程池
            condition_unlock(&pool->ready);
            //执行任务
            t->run(t->arg);
            //执行完任务释放内存
            free(t);
            //重新加锁
            condition_lock(&pool->ready);
        }

        //退出线程池
        if(pool->quit && pool->first == NULL)
        {
            pool->counter--;//当前工作的线程数-1
            //若线程池中没有线程，通知等待线程（主线程）全部任务已经完成
            if(pool->counter == 0)
            {
                condition_signal(&pool->ready);
            }
            condition_unlock(&pool->ready);
            break;
        }
        //超时，跳出销毁线程
        if(timeout == 1)
        {
            pool->counter--;//当前工作的线程数-1
            condition_unlock(&pool->ready);
            break;
        }

        condition_unlock(&pool->ready);
    }

    printf("thread %d is exiting\n", (int)pthread_self());
    return NULL;

}


//线程池初始化
void threadpool_init(threadpool_t *pool, int threads)
{

    condition_init(&pool->ready);
    pool->first = NULL;
    pool->last =NULL;
    pool->counter =0;
    pool->idle =0;
    pool->max_threads = threads;
    pool->quit =0;

}


//增加一个任务到线程池
void threadpool_add_task(threadpool_t *pool, void *(*run)(void *arg), void *arg)
{
    //产生一个新的任务
    task_t *newtask = (task_t *)malloc(sizeof(task_t));
    newtask->run = run;
    newtask->arg = arg;
    newtask->next=NULL;//新加的任务放在队列尾端

    //线程池的状态被多个线程共享，操作前需要加锁
    condition_lock(&pool->ready);

    if(pool->first == NULL)//第一个任务加入
    {
        pool->first = newtask;
    }        
    else    
    {
        pool->last->next = newtask;
    }
    pool->last = newtask;  //队列尾指向新加入的线程

    //线程池中有线程空闲，唤醒
    if(pool->idle > 0)
    {
        condition_signal(&pool->ready);
    }
    //当前线程池中线程个数没有达到设定的最大值，创建一个新的线性
    else if(pool->counter < pool->max_threads)
    {
        pthread_t tid;
        pthread_create(&tid, NULL, thread_routine, pool);
        pool->counter++;
    }
    //结束，访问
    condition_unlock(&pool->ready);
}

//线程池销毁
void threadpool_destroy(threadpool_t *pool)
{
    //如果已经调用销毁，直接返回
    if(pool->quit)
    {
    return;
    }
    //加锁
    condition_lock(&pool->ready);
    //设置销毁标记为1
    pool->quit = 1;
    //线程池中线程个数大于0
    if(pool->counter > 0)
    {
        //对于等待的线程，发送信号唤醒
        if(pool->idle > 0)
        {
            condition_broadcast(&pool->ready);
        }
        //正在执行任务的线程，等待他们结束任务
        while(pool->counter)
        {
            condition_wait(&pool->ready);
        }
    }
    condition_unlock(&pool->ready);
    condition_destroy(&pool->ready);
}