摄像头标定

标定原理介绍

摄像机小孔模型
Cv照相机定标和三维重建#针孔相机模型和变形

[编辑]

标定程序1（opencv自带的示例程序）

[编辑]

简介

读者可以直接使用Opencv自带的摄像机标定示例程序，该程序位于 “\OpenCV\samples\c目录下的calibration.cpp”，程序的输入支持直接从USB摄像机读取图片标定，或者读取avi文件或者已经存放于电脑上图片进行标定。

[编辑]

使用说明

编译运行程序，如果未设置任何命令行参数，则程序会有提示，告诉你应该在你编译出来的程序添加必要的命令行，比如你的程序是calibration.exe(以windows操作系统为例）。则你可以添加如下命令行（以下加粗的字体所示）：

calibration -w 6 -h 8 -s 2 -n 10 -o camera.yml -op -oe [<list_of_views.txt>]

[编辑]

调用命令行和参数介绍

Usage: calibration

-w <board_width>         # 图片某一维方向上的交点个数
-h <board_height>        # 图片另一维上的交点个数
[-n <number_of_frames>]  # 标定用的图片帧数
# (if not specified, it will be set to the number
#  of board views actually available)
[-d <delay>]             # a minimum delay in ms between subsequent attempts to capture a next view
# (used only for video capturing)
[-s <square_size>]       # square size in some user-defined units (1 by default)
[-o <out_camera_params>] # the output filename for intrinsic [and extrinsic] parameters
[-op]                    # write detected feature points
[-oe]                    # write extrinsic parameters
[-zt]                    # assume zero tangential distortion
[-a <aspect_ratio>]      # fix aspect ratio (fx/fy)
[-p]                     # fix the principal point at the center
[-v]                     # flip the captured images around the horizontal axis
[input_data]             # 输入数据，是下面三种之中的一种:
#  - 指定的包含图片列表的txt文件
#  - name of video file with a video of the board
# if input_data not specified, a live view from the camera is used

标定图片示例

上图中，横向和纵向分别为9个交点和6个交点，对应上面的命令行的命令参数应该为： -w 9 -h 6。

经多次使用发现，不指定 -p参数时计算的结果误差较大，主要表现在对u0,v0的估计误差较大，因此建议使用时加上-p参数

[编辑]

list_of_views.txt

该txt文件表示的是你在电脑上面需要用以标定的图片列表。

view000.png
view001.png
#view002.png
view003.png
view010.png
one_extra_view.jpg

上面的例子中，前面加“井号”的图片被忽略。

在windows的命令行中，有一种简便的办法来产生此txt文件。在CMD窗口中输入如下命令（假设当前目录里面的所有jpg文件都用作标定,并且生成的文件为a.txt）。

dir *.jpg /B >> a.txt

[编辑]

输入为摄像机或者avi文件时

"When the live video from camera is used as input, the following hot-keys may be used:\n"
"  <ESC>, 'q' - quit the program\n"
"  'g' - start capturing images\n"
"  'u' - switch undistortion on/off\n";

[编辑]

代码

请直接复制 calibration.cpp 中的相关代码。

[编辑]

标定程序2

OPENCV没有提供完整的示例，自己整理了一下，贴出来记录。

首先自制一张标定图片，用A4纸打印出来，设定距离，再设定标定棋盘的格子数目，如8×6，以下是我做的图片8×8





然后利用cvFindChessboardCorners找到棋盘在摄像头中的2D位置，这里cvFindChessboardCorners不太稳定，有时不能工作，也许需要图像增强处理。

计算实际的距离，应该是3D的距离。我设定为21.6毫米，既在A4纸上为两厘米。
再用cvCalibrateCamera2计算内参，
最后用cvUndistort2纠正图像的变形。

结果如下：





[编辑]

代码

代码下载

具体的函数使用，请参考Cv照相机定标和三维重建#照相机定标

#include "stdafx.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
// OpenCV
#include <cxcore.h>
#include <cv.h>
#include <highgui.h>
#include <cvaux.h>

void InitCorners3D(CvMat *Corners3D, CvSize ChessBoardSize, int Nimages, float SquareSize);
void makeChessBoard();
int myFindChessboardCorners( const void* image, CvSize pattern_size,
CvPoint2D32f* corners, int* corner_count=NULL,
int flags=CV_CALIB_CB_ADAPTIVE_THRESH );

inline int drawCorssMark(IplImage *dst,CvPoint pt)
/*************************************************
Function:        main_loop
Description:     绘制一个十字标记
Calls:
Called By:
Input:           RGB image,  pt
Output:
Return:
Others:          需要检查坐标是否越界 to do list
*************************************************/
{

const int cross_len = 4;
CvPoint pt1,pt2,pt3,pt4;
pt1.x = pt.x;
pt1.y = pt.y - cross_len;
pt2.x = pt.x;
pt2.y = pt.y + cross_len;
pt3.x = pt.x - cross_len;
pt3.y = pt.y;
pt4.x = pt.x + cross_len;
pt4.y = pt.y;

cvLine(dst,pt1,pt2,CV_RGB(0,255,0),2,CV_AA, 0 );
cvLine(dst,pt3,pt4,CV_RGB(0,255,0),2,CV_AA, 0 );

return 0;
}

/* declarations for OpenCV */
IplImage                 *current_frame_rgb,grid;
IplImage                 *current_frame_gray;
IplImage                 *chessBoard_Img;

int                       Thresholdness = 120;

int image_width = 320;
int image_height = 240;

bool verbose = false;

const int ChessBoardSize_w = 7;
const int ChessBoardSize_h = 7;
// Calibration stuff
bool			calibration_done = false;
const CvSize 	ChessBoardSize = cvSize(ChessBoardSize_w,ChessBoardSize_h);
//float 			SquareWidth = 21.6f; //实际距离 毫米单位 在A4纸上为两厘米
float 			SquareWidth = 17; //投影实际距离 毫米单位  200

const   int NPoints = ChessBoardSize_w*ChessBoardSize_h;
const   int NImages = 20; //Number of images to collect

CvPoint2D32f corners[NPoints*NImages];
int corner_count[NImages] = {0};
int captured_frames = 0;

CvMat *intrinsics;
CvMat *distortion_coeff;
CvMat *rotation_vectors;
CvMat *translation_vectors;
CvMat *object_points;
CvMat *point_counts;
CvMat *image_points;
int find_corners_result =0 ;

void on_mouse( int event, int x, int y, int flags, void* param )
{

if( event == CV_EVENT_LBUTTONDOWN )
{
//calibration_done = true;
}
}

int main(int argc, char *argv[])
{

CvFont font;
cvInitFont( &font, CV_FONT_VECTOR0,5, 5, 0, 7, 8);

intrinsics 		= cvCreateMat(3,3,CV_32FC1);
distortion_coeff 	= cvCreateMat(1,4,CV_32FC1);
rotation_vectors 	= cvCreateMat(NImages,3,CV_32FC1);
translation_vectors 	= cvCreateMat(NImages,3,CV_32FC1);

point_counts 		= cvCreateMat(NImages,1,CV_32SC1);

object_points 	= cvCreateMat(NImages*NPoints,3,CV_32FC1);
image_points 		= cvCreateMat(NImages*NPoints,2,CV_32FC1);

// Function to fill in the real-world points of the checkerboard
InitCorners3D(object_points, ChessBoardSize, NImages, SquareWidth);

CvCapture* capture = 0;

if( argc == 1 || (argc == 2 && strlen(argv[1]) == 1 && isdigit(argv[1][0])))
capture = cvCaptureFromCAM( argc == 2 ? argv[1][0] - '0' : 0 );
else if( argc == 2 )
capture = cvCaptureFromAVI( argv[1] );

if( !capture )
{
fprintf(stderr,"Could not initialize capturing...\n");
return -1;
}

// Initialize all of the IplImage structures
current_frame_rgb = cvCreateImage(cvSize(image_width, image_height), IPL_DEPTH_8U, 3);

IplImage *current_frame_rgb2 = cvCreateImage(cvSize(image_width, image_height), IPL_DEPTH_8U, 3);
current_frame_gray = cvCreateImage(cvSize(image_width, image_height), IPL_DEPTH_8U, 1);

chessBoard_Img   = cvCreateImage(cvSize(image_width, image_height), IPL_DEPTH_8U, 3);
current_frame_rgb2->origin = chessBoard_Img->origin  = current_frame_gray->origin = current_frame_rgb->origin = 1;

makeChessBoard();

cvNamedWindow( "result", 0);
cvNamedWindow( "Window 0", 0);
cvNamedWindow( "grid", 0);
cvMoveWindow( "grid", 100,100);
cvSetMouseCallback( "Window 0", on_mouse, 0 );
cvCreateTrackbar("Thresholdness","Window 0",&Thresholdness, 255,0);

while (!calibration_done)
{

while (captured_frames < NImages)
{
current_frame_rgb = cvQueryFrame( capture );
//current_frame_rgb = cvLoadImage( "c:\\BoardStereoL3.jpg" );
//cvCopy(chessBoard_Img,current_frame_rgb);

if( !current_frame_rgb )
break;

cvCopy(current_frame_rgb,current_frame_rgb2);
cvCvtColor(current_frame_rgb, current_frame_gray, CV_BGR2GRAY);
//cvThreshold(current_frame_gray,current_frame_gray,Thresholdness,255,CV_THRESH_BINARY);
//cvThreshold(current_frame_gray,current_frame_gray,150,255,CV_THRESH_BINARY_INV);

/*
int pos = 1;
IplConvKernel* element = 0;
const int element_shape = CV_SHAPE_ELLIPSE;
element = cvCreateStructuringElementEx( pos*2+1, pos*2+1, pos, pos, element_shape, 0 );
cvDilate(current_frame_gray,current_frame_gray,element,1);
cvErode(current_frame_gray,current_frame_gray,element,1);
cvReleaseStructuringElement(&element);
*/

find_corners_result = cvFindChessboardCorners(current_frame_gray,
ChessBoardSize,
&corners[captured_frames*NPoints],
&corner_count[captured_frames],
0);

cvDrawChessboardCorners(current_frame_rgb2, ChessBoardSize, &corners[captured_frames*NPoints], NPoints, find_corners_result);

cvShowImage("Window 0",current_frame_rgb2);
cvShowImage("grid",chessBoard_Img);

if(find_corners_result==1)
{
cvWaitKey(2000);
cvSaveImage("c:\\hardyinCV.jpg",current_frame_rgb2);
captured_frames++;
}
//cvShowImage("result",current_frame_gray);

intrinsics->data.fl[0] = 256.8093262;   //fx
intrinsics->data.fl[2] = 160.2826538;   //cx
intrinsics->data.fl[4] = 254.7511139;   //fy
intrinsics->data.fl[5] = 127.6264572;   //cy

intrinsics->data.fl[1] = 0;
intrinsics->data.fl[3] = 0;
intrinsics->data.fl[6] = 0;
intrinsics->data.fl[7] = 0;
intrinsics->data.fl[8] = 1;

distortion_coeff->data.fl[0] = -0.193740;  //k1
distortion_coeff->data.fl[1] = -0.378588;  //k2
distortion_coeff->data.fl[2] = 0.028980;   //p1
distortion_coeff->data.fl[3] = 0.008136;   //p2

cvWaitKey(40);
find_corners_result = 0;
}
//if (find_corners_result !=0)
{

printf("\n");

cvSetData( image_points, corners, sizeof(CvPoint2D32f));
cvSetData( point_counts, &corner_count, sizeof(int));

cvCalibrateCamera2( object_points,
image_points,
point_counts,
cvSize(image_width,image_height),
intrinsics,
distortion_coeff,
rotation_vectors,
translation_vectors,
0);

// [fx 0 cx; 0 fy cy; 0 0 1].
cvUndistort2(current_frame_rgb,current_frame_rgb,intrinsics,distortion_coeff);
cvShowImage("result",current_frame_rgb);

float intr[3][3] = {0.0};
float dist[4] = {0.0};
float tranv[3] = {0.0};
float rotv[3] = {0.0};

for ( int i = 0; i < 3; i++)
{
for ( int j = 0; j < 3; j++)
{
intr[i][j] = ((float*)(intrinsics->data.ptr + intrinsics->step*i))[j];
}
dist[i] = ((float*)(distortion_coeff->data.ptr))[i];
tranv[i] = ((float*)(translation_vectors->data.ptr))[i];
rotv[i] = ((float*)(rotation_vectors->data.ptr))[i];
}
dist[3] = ((float*)(distortion_coeff->data.ptr))[3];

printf("-----------------------------------------\n");
printf("INTRINSIC MATRIX: \n");
printf("[ %6.4f %6.4f %6.4f ] \n", intr[0][0], intr[0][1], intr[0][2]);
printf("[ %6.4f %6.4f %6.4f ] \n", intr[1][0], intr[1][1], intr[1][2]);
printf("[ %6.4f %6.4f %6.4f ] \n", intr[2][0], intr[2][1], intr[2][2]);
printf("-----------------------------------------\n");
printf("DISTORTION VECTOR: \n");
printf("[ %6.4f %6.4f %6.4f %6.4f ] \n", dist[0], dist[1], dist[2], dist[3]);
printf("-----------------------------------------\n");
printf("ROTATION VECTOR: \n");
printf("[ %6.4f %6.4f %6.4f ] \n", rotv[0], rotv[1], rotv[2]);
printf("TRANSLATION VECTOR: \n");
printf("[ %6.4f %6.4f %6.4f ] \n", tranv[0], tranv[1], tranv[2]);
printf("-----------------------------------------\n");

cvWaitKey(0);

calibration_done = true;
}

}

exit(0);
cvDestroyAllWindows();
}

void InitCorners3D(CvMat *Corners3D, CvSize ChessBoardSize, int NImages, float SquareSize)
{
int CurrentImage = 0;
int CurrentRow = 0;
int CurrentColumn = 0;
int NPoints = ChessBoardSize.height*ChessBoardSize.width;
float * temppoints = new float[NImages*NPoints*3];

// for now, assuming we're row-scanning
for (CurrentImage = 0 ; CurrentImage < NImages ; CurrentImage++)
{
for (CurrentRow = 0; CurrentRow < ChessBoardSize.height; CurrentRow++)
{
for (CurrentColumn = 0; CurrentColumn < ChessBoardSize.width; CurrentColumn++)
{
temppoints[(CurrentImage*NPoints*3)+(CurrentRow*ChessBoardSize.width + CurrentColumn)*3]=(float)CurrentRow*SquareSize;
temppoints[(CurrentImage*NPoints*3)+(CurrentRow*ChessBoardSize.width + CurrentColumn)*3+1]=(float)CurrentColumn*SquareSize;
temppoints[(CurrentImage*NPoints*3)+(CurrentRow*ChessBoardSize.width + CurrentColumn)*3+2]=0.f;
}
}
}
(*Corners3D) = cvMat(NImages*NPoints,3,CV_32FC1, temppoints);
}

int myFindChessboardCorners( const void* image, CvSize pattern_size,
CvPoint2D32f* corners, int* corner_count,
int flags )

{

IplImage* eig = cvCreateImage( cvGetSize(image), 32, 1 );
IplImage* temp = cvCreateImage( cvGetSize(image), 32, 1 );
double quality = 0.01;
double min_distance = 5;
int win_size =10;

int count = pattern_size.width * pattern_size.height;
cvGoodFeaturesToTrack( image, eig, temp, corners, &count,
quality, min_distance, 0, 3, 0, 0.04 );
cvFindCornerSubPix( image, corners, count,
cvSize(win_size,win_size), cvSize(-1,-1),
cvTermCriteria(CV_TERMCRIT_ITER|CV_TERMCRIT_EPS,20,0.03));

cvReleaseImage( &eig );
cvReleaseImage( &temp );

return 1;
}

void makeChessBoard()
{

CvScalar e;
e.val[0] =255;
e.val[1] =255;
e.val[2] =255;
cvSet(chessBoard_Img,e,0);
for(int i = 0;i<ChessBoardSize.width+1;i++)
for(int j = 0;j<ChessBoardSize.height+1;j++)
{
int w =(image_width)/2/(ChessBoardSize.width);
int h = w; //(image_height)/2/(ChessBoardSize.height);

int ii = i+1;
int iii = ii+1;
int jj =j+1;
int jjj =jj+1;
int s_x = image_width/6;

if((i+j)%2==1)
cvRectangle( chessBoard_Img, cvPoint(w*i+s_x,h*j+s_x),cvPoint(w*ii-1+s_x,h*jj-1+s_x), CV_RGB(0,0,0),CV_FILLED, 8, 0 );
}
}