1，适用于VideoCapture打开的摄像头
VideoCapture capture(0); 设置摄像头参数 不要随意修改
capture.set(CV_CAP_PROP_FRAME_WIDTH, 1080);//宽度
capture.set(CV_CAP_PROP_FRAME_HEIGHT, 960);//高度
capture.set(CV_CAP_PROP_FPS, 30);//帧率 帧/秒
capture.set(CV_CAP_PROP_BRIGHTNESS, 1);//亮度
capture.set(CV_CAP_PROP_CONTRAST,40);//对比度 40
capture.set(CV_CAP_PROP_SATURATION, 50);//饱和度 50
capture.set(CV_CAP_PROP_HUE, 50);//色调 50
capture.set(CV_CAP_PROP_EXPOSURE, 50);//曝光 50 获取摄像头参数
得到摄像头的参数
capture.get(CV_CAP_PROP_FRAME_WIDTH);
capture.get(CV_CAP_PROP_FRAME_HEIGHT);
capture.get(CV_CAP_PROP_FPS);
capture.get(CV_CAP_PROP_BRIGHTNESS);
capture.get(CV_CAP_PROP_CONTRAST);
capture.get(CV_CAP_PROP_SATURATION);
capture.get(CV_CAP_PROP_HUE);
capture.get(CV_CAP_PROP_EXPOSURE); 获取视频参数：
capture.get(CV_CAP_PROP_FRAME_COUNT);//视频帧数
然后你会发现除了个别参数你能更改之外（如曝光度），大分布你是不能更改的，甚至都没办法得到，这种并不适用
2，不做开发，只是单纯的更改
那么推荐一个软件，amcap，百度网盘链接，
https://pan.baidu.com/s/1pL8nq0V#list/path=%2F
，很简单很容易上手。
补，现在突然想起来我的一个学长告诉我的，利用这个软件调节摄像头的曝光度，可以改变帧率，且摄像头会记住曝光度的设置（其他特性就没有这个特点）。-2019.3.12
3，修改opencv的文件，不过效果可能和第一个差不多
大概是在opencv的这个位置，找一下，modules/highgui/src/cap_v4l.cpp，里面有关于参数的设置，位置比较靠前，可以搜索，也可以直接找到
大致在200多行
4，v4l2
下面是我找到的一篇参考，可以突破帧率的限制，当然前提是摄像头支持
https://blog.csdn.net/c406495762/article/details/72732135
目前只适用于Linux系统，本人试验过，120帧的摄像头在只打开摄像头时可以达到100帧左右，设置的图片分辨率越小，能达到的帧率越高
#include <unistd.h>
#include <error.h>
#include <errno.h>
#include <fcntl.h>
#include <sys/ioctl.h>
#include <sys/types.h>
#include <pthread.h>
#include <linux/videodev2.h>
#include <sys/mman.h>
#include <opencv2/core/core.hpp>
#include <opencv2/highgui/highgui.hpp>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <time.h>
#include "opencv2/highgui/highgui.hpp"
#include "opencv2/imgproc/imgproc.hpp"
#include <math.h>
#include <iostream>
#include <iomanip>
#include <string>
using namespace std;
using namespace cv;
#define CLEAR(x) memset(&(x), 0, sizeof(x))
#define IMAGEWIDTH 3264
#define IMAGEHEIGHT 2448
#define WINDOW_NAME1 "【原始图】" //为窗口标题定义的宏
#define WINDOW_NAME2 "【图像轮廓】" //为窗口标题定义的宏
Mat g_srcImage; Mat g_grayImage;
int g_nThresh = 90;
int g_nMaxThresh = 255;
RNG g_rng(12345);
Mat g_cannyMat_output;
vector<vector<Point> > g_vContours;
vector<Vec4i> g_vHierarchy;
Point point1[100000];
Point point2[100000];
Point point3[100000];
int ii,iii;
int flag2 = 0;//避障用
float number = 0;
int fps=0;
class V4L2Capture {
public:
V4L2Capture(char *devName, int width, int height);
virtual ~V4L2Capture();
int openDevice();
int closeDevice();
int initDevice();
int startCapture();
int stopCapture();
int freeBuffers();
int getFrame(void **,size_t *);
int backFrame();
static void test();
private:
int initBuffers();
struct cam_buffer
void* start;
unsigned int length;
char *devName;
int capW;
int capH;
int fd_cam;
cam_buffer *buffers;
unsigned int n_buffers;
int frameIndex;
V4L2Capture::V4L2Capture(char *devName, int width, int height) {
// TODO Auto-generated constructor stub
this->devName = devName;
this->fd_cam = -1;
this->buffers = NULL;
this->n_buffers = 0;
this->frameIndex = -1;
this->capW=width;
this->capH=height;
V4L2Capture::~V4L2Capture() {
// TODO Auto-generated destructor stub
int V4L2Capture::openDevice() {
/*设备的打开*/
printf("video dev : %s\n", devName);
fd_cam = open(devName, O_RDWR);
if (fd_cam < 0) {
perror("Can't open video device");
return 0;
int V4L2Capture::closeDevice() {
if (fd_cam > 0) {
int ret = 0;
if ((ret = close(fd_cam)) < 0) {
perror("Can't close video device");
return 0;
} else {
return -1;
int V4L2Capture::initDevice() {
int ret;
struct v4l2_capability cam_cap; //显示设备信息
struct v4l2_cropcap cam_cropcap; //设置摄像头的捕捉能力
struct v4l2_fmtdesc cam_fmtdesc; //查询所有支持的格式：VIDIOC_ENUM_FMT
struct v4l2_crop cam_crop; //图像的缩放
struct v4l2_format cam_format; //设置摄像头的视频制式、帧格式等
/* 使用IOCTL命令VIDIOC_QUERYCAP，获取摄像头的基本信息*/
ret = ioctl(fd_cam, VIDIOC_QUERYCAP, &cam_cap);
if (ret < 0) {
perror("Can't get device information: VIDIOCGCAP");
printf(
"Driver Name:%s\nCard Name:%s\nBus info:%s\nDriver Version:%u.%u.%u\n",
cam_cap.driver, cam_cap.card, cam_cap.bus_info,
(cam_cap.version >> 16) & 0XFF, (cam_cap.version >> 8) & 0XFF,
cam_cap.version & 0XFF);
/* 使用IOCTL命令VIDIOC_ENUM_FMT，获取摄像头所有支持的格式*/
cam_fmtdesc.index = 0;
cam_fmtdesc.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
printf("Support format:\n");
while (ioctl(fd_cam, VIDIOC_ENUM_FMT, &cam_fmtdesc) != -1) {
printf("\t%d.%s\n", cam_fmtdesc.index + 1, cam_fmtdesc.description);
cam_fmtdesc.index++;
/* 使用IOCTL命令VIDIOC_CROPCAP，获取摄像头的捕捉能力*/
cam_cropcap.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
if (0 == ioctl(fd_cam, VIDIOC_CROPCAP, &cam_cropcap)) {
printf("Default rec:\n\tleft:%d\n\ttop:%d\n\twidth:%d\n\theight:%d\n",
cam_cropcap.defrect.left, cam_cropcap.defrect.top,
cam_cropcap.defrect.width, cam_cropcap.defrect.height);
/* 使用IOCTL命令VIDIOC_S_CROP，获取摄像头的窗口取景参数*/
cam_crop.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
cam_crop.c = cam_cropcap.defrect; //默认取景窗口大小
if (-1 == ioctl(fd_cam, VIDIOC_S_CROP, &cam_crop)) {
//printf("Can't set crop para\n");
} else {
printf("Can't set cropcap para\n");
/* 使用IOCTL命令VIDIOC_S_FMT，设置摄像头帧信息*/
cam_format.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
cam_format.fmt.pix.width = capW;
cam_format.fmt.pix.height = capH;
cam_format.fmt.pix.pixelformat = V4L2_PIX_FMT_MJPEG; //要和摄像头支持的类型对应
cam_format.fmt.pix.field = V4L2_FIELD_INTERLACED;
ret = ioctl(fd_cam, VIDIOC_S_FMT, &cam_format);
if (ret < 0) {
perror("Can't set frame information");
/* 使用IOCTL命令VIDIOC_G_FMT，获取摄像头帧信息*/
cam_format.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
ret = ioctl(fd_cam, VIDIOC_G_FMT, &cam_format);
if (ret < 0) {
perror("Can't get frame information");
printf("Current data format information:\n\twidth:%d\n\theight:%d\n",
cam_format.fmt.pix.width, cam_format.fmt.pix.height);
ret = initBuffers();
if (ret < 0) {
perror("Buffers init error");
//exit(-1);
return 0;
int V4L2Capture::initBuffers() {
int ret;
/* 使用IOCTL命令VIDIOC_REQBUFS，申请帧缓冲*/
struct v4l2_requestbuffers req;
CLEAR(req);
req.count = 4;
req.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
req.memory = V4L2_MEMORY_MMAP;
ret = ioctl(fd_cam, VIDIOC_REQBUFS, &req);
if (ret < 0) {
perror("Request frame buffers failed");
if (req.count < 2) {
perror("Request frame buffers while insufficient buffer memory");
buffers = (struct cam_buffer*) calloc(req.count, sizeof(*buffers));
if (!buffers) {
perror("Out of memory");
for (n_buffers = 0; n_buffers < req.count; n_buffers++) {
struct v4l2_buffer buf;
CLEAR(buf);
// 查询序号为n_buffers 的缓冲区，得到其起始物理地址和大小
buf.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
buf.memory = V4L2_MEMORY_MMAP;
buf.index = n_buffers;
ret = ioctl(fd_cam, VIDIOC_QUERYBUF, &buf);
if (ret < 0) {
printf("VIDIOC_QUERYBUF %d failed\n", n_buffers);
return -1;
buffers[n_buffers].length = buf.length;
//printf("buf.length= %d\n",buf.length);
// 映射内存
buffers[n_buffers].start = mmap(
NULL, // start anywhere
buf.length, PROT_READ | PROT_WRITE, MAP_SHARED, fd_cam,
buf.m.offset);
if (MAP_FAILED == buffers[n_buffers].start) {
printf("mmap buffer%d failed\n", n_buffers);
return -1;
return 0;
int V4L2Capture::startCapture() {
unsigned int i;
for (i = 0; i < n_buffers; i++) {
struct v4l2_buffer buf;
CLEAR(buf);
buf.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
buf.memory = V4L2_MEMORY_MMAP;
buf.index = i;
if (-1 == ioctl(fd_cam, VIDIOC_QBUF, &buf)) {
printf("VIDIOC_QBUF buffer%d failed\n", i);
return -1;
enum v4l2_buf_type type;
type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
if (-1 == ioctl(fd_cam, VIDIOC_STREAMON, &type)) {
printf("VIDIOC_STREAMON error");
return -1;
return 0;
int V4L2Capture::stopCapture() {
enum v4l2_buf_type type;
type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
if (-1 == ioctl(fd_cam, VIDIOC_STREAMOFF, &type)) {
printf("VIDIOC_STREAMOFF error\n");
return -1;
return 0;
}/*ok*/
int V4L2Capture::freeBuffers() {
unsigned int i;
for (i = 0; i < n_buffers; ++i) {
if (-1 == munmap(buffers[i].start, buffers[i].length)) {
printf("munmap buffer%d failed\n", i);
return -1;
free(buffers);
return 0;
int V4L2Capture::getFrame(void **frame_buf, size_t* len) {
struct v4l2_buffer queue_buf;
CLEAR(queue_buf);
queue_buf.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
queue_buf.memory = V4L2_MEMORY_MMAP;
if (-1 == ioctl(fd_cam, VIDIOC_DQBUF, &queue_buf)) {
printf("VIDIOC_DQBUF error\n");
return -1;
*frame_buf = buffers[queue_buf.index].start;
*len = buffers[queue_buf.index].length;
frameIndex = queue_buf.index;
return 0;
int V4L2Capture::backFrame() {
if (frameIndex != -1) {
struct v4l2_buffer queue_buf;
CLEAR(queue_buf);
queue_buf.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
queue_buf.memory = V4L2_MEMORY_MMAP;
queue_buf.index = frameIndex;
if (-1 == ioctl(fd_cam, VIDIOC_QBUF, &queue_buf)) {
printf("VIDIOC_QBUF error\n");
return -1;
return 0;
return -1;
void V4L2Capture::test() {
unsigned char *yuv422frame = NULL;
unsigned long yuvframeSize = 0;
string videoDev="/dev/video0";
V4L2Capture *vcap = new V4L2Capture(const_cast<char*>(videoDev.c_str()),
1920, 1080);
vcap->openDevice();
vcap->initDevice();
vcap->startCapture();
vcap->getFrame((void **) &yuv422frame, (size_t *)&yuvframeSize);
vcap->backFrame();
vcap->freeBuffers();
vcap->closeDevice();
void line2(Point point3[100000], int n)
float aa, bb, cc, dd, ee, ff, gg;
int jj = 0;
for (;jj <n;jj++)
aa += point3[jj].x*point3[jj].x;
bb += point3[jj].x;
cc += point3[jj].x*point3[jj].y;
dd += point3[jj].y;
ee = aa*n - bb*bb;
if ((int)(ee* 100) != 0)
ff = (n*cc - bb*dd) / ee;
gg = (dd - bb*ff) / n;
else {
ff = 0;
gg = 1;
Point point0, pointn;
point0.y = 0;
point0.x = gg;
pointn.y = (n-1);
pointn.x = ((n-1) * ff + gg);
Mat draw_ing2 = Mat::zeros(g_cannyMat_output.size(), CV_8UC3);
line(draw_ing2, point0, pointn, (255, 255, 255));
imshow("10", draw_ing2);
//cout << "\n"<<ff <<" "<< gg << endl;
float the =180*atan(ff)/3.14159;
float dis = ff * 160+gg - 160;
cout << the << " " << dis << endl;
//正中心ff=0，gg=160，逆时ff为正，顺时ff为负
void findcolor(cv::Mat &image)
cv::Mat_<cv::Vec3b>::iterator it = image.begin<cv::Vec3b>();
cv::Mat_<cv::Vec3b>::iterator itend = image.end<cv::Vec3b>();
ii = 0;
iii = 0;
int flagg = 0;
cv::Mat srcX(image.rows, image.cols , CV_32F);
cv::Mat srcY(image.rows, image.cols, CV_32F);
for (int i = 0;i < image.rows;i++)
for (int j = 0;j < image.cols;j++)
if (flagg == 0)/*这样遍历水平方向无法得到有效数据*/
if ((*it)[0] == 255 && (*it)[1] == 0 && (*it)[2] == 255)
flagg = 1;
point1[ii].x = i;
point1[ii].y = j;
ii++;
if ((*it)[0] == 255 && (*it)[1] == 0 && (*it)[2] == 255)
flagg = 0;
point2[iii].x = i;
point2[iii].y = j;
iii++;
if (it == itend)
break;
else it++;
IplImage pImg = IplImage(image);
CvArr* arr = (CvArr*)&pImg;
int nn = ii;
for (;ii > 0;ii--)
point3[ii].x = (point1[ii].x + point2[ii].x) / 2;
point3[ii].y = (point1[ii].y + point2[ii].y) / 2;
//circle(image, point3[ii], 1, (255, 255, 255));
cvSet2D(arr, point3[ii].x, point3[ii].y, Scalar(255, 255, 255));
line2(point3, nn);
void on_ThreshChange(int, void* )
// 使用Canndy检测边缘
Canny( g_grayImage, g_cannyMat_output, g_nThresh, g_nThresh*2, 3 );
// 找到轮廓
findContours( g_cannyMat_output, g_vContours, g_vHierarchy, RETR_TREE, CHAIN_APPROX_SIMPLE, Point(0, 0) );
// 计算矩
vector<Moments> mu(g_vContours.size() );
for(unsigned int i = 0; i < g_vContours.size(); i++ )
{ mu[i] = moments( g_vContours[i], false ); }
// 计算中心矩
vector<Point2f> mc( g_vContours.size() );
for( unsigned int i = 0; i < g_vContours.size(); i++ )
{ mc[i] = Point2f( static_cast<float>(mu[i].m10/mu[i].m00), static_cast<float>(mu[i].m01/mu[i].m00 )); }
// 绘制轮廓
Mat drawing = Mat::zeros(g_cannyMat_output.size(), CV_8UC3);
for( unsigned int i = 0; i< g_vContours.size(); i++ )
//Scalar color = Scalar( g_rng.uniform(0, 255), g_rng.uniform(0,255), g_rng.uniform(0,255) );//随机生成颜色值
Scalar color = Scalar(255, 0, 255);
drawContours( drawing, g_vContours, i, color, 2, 8, g_vHierarchy, 0, Point() );//绘制外层和内层轮廓
circle( drawing, mc[i], 4, color, -1, 8, 0 );;//绘制圆
findcolor(drawing);
//line1(point1,point2,ii,iii);
// 显示到窗口中
// namedWindow( WINDOW_NAME2, WINDOW_AUTOSIZE );
imshow( WINDOW_NAME2, drawing );
void findline(Mat image)
cv::Mat_<cv::Vec3b>::iterator it = image.begin<cv::Vec3b>();
cv::Mat_<cv::Vec3b>::iterator itend = image.end<cv::Vec3b>();
for (;it != itend;it++)
if ((*it)[1] == 0 && (*it)[2] >= 100)//条件可能需要改变
if(flag2==0)
flag2 = 1;
cout << "注意line1，避障"<<endl;
//向主控发送消息
cout << "注意line2，避障" << endl;
//向主控发送消息
//避障一与避障二中间要隔一段时间
void wave(const cv::Mat &image, cv::Mat &result)
cv::Mat srcX(image.rows / 2, image.cols / 2, CV_32F);
cv::Mat srcY(image.rows / 2, image.cols / 2, CV_32F);
for (int i = 0;i<image.rows /2;i++)
for (int j = 0;j < image.cols /2;j++)
srcX.at<float>(i, j) = 2 * j;
srcY.at<float>(i, j) = 2 * i;
cv::remap(image, result, srcX, srcY, cv::INTER_LINEAR);
void VideoPlayer() {
unsigned char *yuv422frame = NULL;
unsigned long yuvframeSize = 0;
string videoDev = "/dev/video0";
V4L2Capture *vcap = new V4L2Capture(const_cast<char*>(videoDev.c_str()), 640, 480);
vcap->openDevice();
vcap->initDevice();
vcap->startCapture();
cvNamedWindow("Capture",CV_WINDOW_AUTOSIZE);
IplImage* img;
CvMat cvmat;
double t;
clock_t start, end;
double number=0;
int fps=0;
while(1){
start=clock();
t = (double)cvGetTickCount();
vcap->getFrame((void **) &yuv422frame, (size_t *)&yuvframeSize);
cvmat = cvMat(IMAGEHEIGHT,IMAGEWIDTH,CV_8UC3,(void*)yuv422frame); //CV_8UC3
img = cvDecodeImage(&cvmat,1);
if(!img){
printf("DecodeImage error!\n");
cv::Mat g_srcImage = cv::cvarrToMat(img,true);
cvShowImage("Capture",img);
cvReleaseImage(&img);
vcap->backFrame();
if((cvWaitKey(1)&255) == 27){
exit(0);
wave(g_srcImage, g_srcImage);
findline(g_srcImage);
// 把原图像转化成灰度图像并进行平滑
cvtColor(g_srcImage, g_grayImage, COLOR_BGR2GRAY);
blur(g_grayImage, g_grayImage, Size(3, 3));
//创建滚动条并进行初始化
createTrackbar(" 阈值", WINDOW_NAME1, &g_nThresh, g_nMaxThresh, on_ThreshChange);
on_ThreshChange(0, 0);
t = (double)cvGetTickCount() - t;
printf("Used time is %g ms\n", (t / (cvGetTickFrequency() * 1000)));
end =clock();
number=number+end-start;
fps++;
if (number/ CLOCKS_PER_SEC>= 0.25)//windows10 for CLK_TCK
cout<<fps<<endl;
fps = 0;
number = 0;
vcap->stopCapture();
vcap->freeBuffers();
vcap->closeDevice();
int main() {
VideoPlayer();
return 0;
1，适用于VideoCapture打开的摄像头VideoCapture capture(0); 设置摄像头参数不要随意修改capture.set(CV_CAP_PROP_FRAME_WIDTH, 1080);//宽度capture.set(CV_CAP_PROP_FRAME_HEIGHT, 960);//高度capture.set(CV_CAP_PROP_FPS, 30);//帧...
1，为了获取视频,你应该创建一个 VideoCapture 对象。他的参数可以是设备的索引号,或者是一个视频文件。设备索引号就是在指定要使用的摄像头。一般的笔记本电脑都有内置摄像头。所以参数就是 0。你可以通过设置成 1 或者其他的来选择别的摄像头。之后,你就可以一帧一帧的捕获视频了。但是最后,别忘了停止捕获视频。使用 ls /dev/video*命令可以查看摄像头设备
2，cap.read() 返回一个布尔值(True/False)。如果帧读取的是正确的,就是 True。所以最后你可以通过检查他的返回值来查看视频文件是否已经到了结尾。有时 cap 可能不能成功的初始化摄像头设备。这种情况下上
之前找了好久Opencv与DirectShow结合使用的程序，甚至动用了2009年大神的代码，今天发现Opencv自己已经可以支持DirectSHOW了，使用VideoCapture类
代码如下：
By longxiaoyue wunnnn@hotmail.com
使用directshow，可以读取100fps640*480，MJPG压缩视频,但是不能加上显示，否则只能到60帧...
2.opencv配置
比如博客安装的opencv路径如下：E:\E05_C++Develop\E05_01_opencv_3.4.5\install\
-a.将dll拷贝至源文件的目录下（项目目录下）
比如拷贝E:\E05_C++Develop\E05_01_opencv_3.4.5\install\opencv\build\x64\vc14\bin\opencv_world345.dll
E:\E05_C
public:
/**1. @brief Default constructor
@note In @ref videoio_c "C API", when you finished working with video, release CvCapture structure with
