用ubuntu想在写代码时，随时观察自己身后的状态，用cheese做录像太明显了，而且还得动鼠标，打算写个用的摄像头拍照的工具。用v4l2应该最容易。
头文件，包括mem，dev和ioctl相关的头文件
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/ioctl.h>
#include <unistd.h>
#include <errno.h>
#include <sys/mman.h>
#include <linux/videodev2.h>
定义一个方便传递的结构体
struct buffer
{
void* start;
unsigned int length;
};
所有的函数定义：
int main()
{
FILE* fp;
int fd_video;
struct buffer *buffers;
//打开摄像头设备
fd_video = open("/dev/video0", O_RDWR);
if (fd_video < 0)
{
perror("video capture open");
return fd_video;
}
fp = fopen("/root/test.jpg", "wb+");
if (fp < 0)
{
perror("fb open error.");
return -1;
}
//打印支持的视频格式
struct v4l2_fmtdesc fmtdesc;
fmtdesc.index=0;
fmtdesc.type=V4L2_BUF_TYPE_VIDEO_CAPTURE;
printf("Support format:\n");
while(ioctl(fd_video, VIDIOC_ENUM_FMT, &fmtdesc) != -1)
{
printf("\t%d.%s ,fmtdesc:%d\n",fmtdesc.index+1,fmtdesc.description,fmtdesc.pixelformat);
fmtdesc.index++;
}
//设置视频的格式，720P JPEG
struct v4l2_format s_fmt;
s_fmt.fmt.pix.width = 1280;
s_fmt.fmt.pix.height = 720;
s_fmt.fmt.pix.pixelformat = V4L2_PIX_FMT_MJPEG;
printf("s_fmt.fmt.pix.pixelformat:%d\n",s_fmt.fmt.pix.pixelformat);
s_fmt.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
int flag= ioctl(fd_video,VIDIOC_S_FMT,&s_fmt);
if(flag != 0)
{
printf("set format error\n");
return -1;
}
printf("jpeg:%d %d\n",s_fmt.fmt.pix.width,s_fmt.fmt.pix.height);
//申请1个缓冲区
struct v4l2_requestbuffers req;
req.count=1;
req.type=V4L2_BUF_TYPE_VIDEO_CAPTURE;
req.memory=V4L2_MEMORY_MMAP;
ioctl(fd_video,VIDIOC_REQBUFS,&req);
//缓冲区与应用程序关联
//申请1个struct buffer空间
buffers = (struct buffer*)calloc (req.count, sizeof (struct buffer));
if (!buffers)
{
perror ("Out of memory");
exit (EXIT_FAILURE);
}
unsigned int n_buffers;
for (n_buffers = 0; n_buffers < req.count; n_buffers++)
{
struct v4l2_buffer buf;
memset(&buf,0,sizeof(buf));
buf.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
buf.memory = V4L2_MEMORY_MMAP;
buf.index = n_buffers;
if (-1 == ioctl (fd_video, VIDIOC_QUERYBUF, &buf))
exit(-1);
buffers[n_buffers].length = buf.length;
buffers[n_buffers].start = mmap (NULL,
buf.length,PROT_READ | PROT_WRITE ,MAP_SHARED,fd_video, buf.m.offset);
if (MAP_FAILED == buffers[n_buffers].start)
exit(-1);
}
enum v4l2_buf_type type;
for (n_buffers = 0; n_buffers < req.count; n_buffers++)
{
struct v4l2_buffer buf;
buf.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
buf.memory = V4L2_MEMORY_MMAP;
buf.index = n_buffers;
ioctl (fd_video, VIDIOC_QBUF, &buf);
}
//开始捕获图像
type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
ioctl (fd_video, VIDIOC_STREAMON, &type);
struct v4l2_buffer buf;
memset(&(buf), 0, sizeof(buf));
buf.type = V4L2_BUF_TYPE_VIDEO_CAPTURE;
buf.memory = V4L2_MEMORY_MMAP;
//取出图像数据
ioctl (fd_video, VIDIOC_DQBUF, &buf);
//保存图像
fwrite(buffers[buf.index].start,1,buffers[buf.index].length,fp);//mjpeg
fflush(fp);
//放回缓冲区
ioctl (fd_video,VIDIOC_QBUF,&buf);
for (n_buffers = 0; n_buffers < req.count; n_buffers++)
munmap(buffers[n_buffers].start, buffers[n_buffers].length);
free(buffers);
close(fd_video);
fclose(fp);
printf("capture jpg finish..\n");
system("eog -w /root/test.jpg &");
}
说明：
打开摄像头和图片
fd_video = open("/dev/video0", O_RDWR);
fp = fopen("/root/test.jpg", "wb+");
查询摄像头支持的格式，我的电脑支持yuyv和mjpeg
ioctl(fd_video, VIDIOC_ENUM_FMT, &fmtdesc)
设置摄像头拍摄的格式，尺寸是720P，格式是MJPEG
ioctl(fd_video,VIDIOC_S_FMT,&s_fmt)
申请缓冲区，一张图片就需要一帧的缓冲区
ioctl(fd_video,VIDIOC_REQBUFS,&req);
用calloc是申请多个缓冲区用的，这里申请1个缓冲区，与malloc一样，这里是在摄像基础上改的，用calloc就不改了。
buffers = (struct buffer*)calloc (req.count, sizeof (struct buffer));
用mmap映射内存的目的，是省去了memcpy的花销，更加合理
buffers[n_buffers].start = mmap (NULL,
buf.length,PROT_READ | PROT_WRITE ,MAP_SHARED,fd_video, buf.m.offset);
设置缓冲区，并放回缓冲区
ioctl (fd_video, VIDIOC_QBUF, &buf)
开始捕获图像
ioctl (fd_video, VIDIOC_STREAMON, &type)
从缓冲区取出一帧数据
ioctl (fd_video, VIDIOC_DQBUF, &buf)
取完数据，再放回缓冲区
ioctl (fd_video, VIDIOC_QBUF, &buf)
之后释放资源，就不赘述了。
通过eye of gnome将jpg显示出来。
system("eog -w /root/test.jpg &");