原本网上可以搜到很多关于双目相机标定的文章，但是一直没得搜到一个完整的从图像获取，再到单目标定，最后到双目标定的全过程。为此，博主写下了这篇博客，尽微薄之力，为大家呈上Help。多言不虚了，上代码》》》》》》》》》》》
1. OpenCV代码之左右图像获取：OpenCV3.0 + VS2010 + 64位操作系统
#include "stdafx.h"
#include <opencv2/opencv.hpp>
#include <iostream>
using namespace std;
using namespace cv;
int main()
{
int i=0;
VideoCapture cap_left(1);
VideoCapture cap_right(0);
if (cap_left.isOpened() && cap_right.isOpened())
cout << "camera is opened" << endl;
else
cout << "camera is not opened" << endl;
bool stop = false;
Mat frame;
cap_left.set(CAP_PROP_FRAME_WIDTH, 480);
cap_left.set(CAP_PROP_FRAME_HEIGHT, 640);
cap_right.set(CAP_PROP_FRAME_WIDTH, 480);
cap_right.set(CAP_PROP_FRAME_HEIGHT, 640);
VideoWriter writer0("two_test0.avi",VideoWriter::fourcc('M', 'J', 'P', 'G'),30, Size(640,480),true);
VideoWriter writer1("two_test1.avi",VideoWriter::fourcc('M', 'J', 'P', 'G'),30, Size(640,480),true);
Mat img0 , img1;
namedWindow("camera_left");
namedWindow("camera_right");
int count = 130;
while (count)
{
cap_left >> img0; //
if(!cap_left.read(img0)) // 读取下一帧图像
break;
writer0 << img0;
imshow("camera_left", img0);
cap_right >> img1;
if(!cap_right.read(img1)) // 读取下一帧图像
break;
writer1 << img1;
imshow("camera_right", img1);
waitKey(30);
count --;
}
//读取获取到的视频，进行图像采集
VideoCapture cap00("two_test0.avi");
VideoCapture cap11("two_test1.avi");
if(cap00.isOpened() && cap11.isOpened())
cout << "ok ,have a video" << endl;
else
cout << "no video" << endl;
Mat frame00, frame11;
int num0 = 1, num1 = 1;
string filename0;
char temp_file0[100];
string path0 = "C:\\Users\\Administrator\\Desktop\\标定图像获取\\left\\left";//can change the path
for (int i = 1; i < 131; i++)
{
cap00 >> frame00;
if(frame00.empty())
break;
if(i % 5 == 0)
{
_itoa_s(num0, temp_file0, 4, 10);
filename0 = temp_file0;
filename0 = path0 + filename0 + ".jpg";
num0 ++;
imwrite (filename0,frame00);
}
}
cap00.release();
//重复上一步骤
string filename1;
char temp_file1[100];
string path1 = "C:\\Users\\Administrator\\Desktop\\标定图像获取\\right\\right";//can change the path
for (int i = 1; i < 131;i++)
{
cap11 >> frame11;
if(frame11.empty())
break;
if(i % 5 == 0)
{
_itoa_s(num1, temp_file1, 4, 10);
filename1 = temp_file1;
filename1 = path1 + filename1 + ".jpg";
num1 ++;
imwrite (filename1,frame11);
}
}
cap11.release();
return 0;
}
Camera Calibration and 3D Reconstruction 相机标定的函数及公式
Left图像：
Image names—>Read images—> Extract grid corners
Calibration按钮
Initialization of the intrinsic parameters - Number of images: 20
Calibration parameters after initialization:
Focal Length: fc = [ 1031.50439 1031.50439 ]
Principal point: cc = [ 319.50000 239.50000 ]
Skew: alpha_c = [ 0.00000 ] => angle of pixel = 90.00000 degrees
Distortion: kc = [ 0.00000 0.00000 0.00000 0.00000 0.00000 ]
Main calibration optimization procedure - Number of images: 20
Gradient descent iterations: 1...2...3...4...5...6...7...8...9...10...11...12...13...14...15...16...17...18...19...20...21...22...23...24...25...26...27...28...29...30...done
Estimation of uncertainties...done
Calibration results after optimization (with uncertainties):
Focal Length: fc = [ 997.33556 1001.73912 ] +/- [ 55.05614 55.49485 ]
Principal point: cc = [ 298.01423 249.24570 ] +/- [ 13.54904 7.56469 ]
Skew: alpha_c = [ 0.00000 ] +/- [ 0.00000 ] => angle of pixel axes = 90.00000 +/- 0.00000 degrees
Distortion: kc = [ 0.01808 1.16917 0.00352 -0.00714 0.00000 ] +/- [ 0.04498 0.74714 0.00278 0.00346 0.00000 ]
Pixel error: err = [ 0.18524 0.15456 ]
Note: The numerical errors are approximately three times the standard deviations (for reference).
Recommendation: Some distortion coefficients are found equal to zero (within their uncertainties).
To reject them from the optimization set est_dist=[0;1;1;1;0] and run Calibration
Show Extrinsic
Analyse error
Recomp.corners ——> Calibration
There is now a total of 20 active images for calibration:
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 and 20.
You may now run 'Calibration' to recalibrate based on this new set of images.
Aspect ratio optimized (est_aspect_ratio = 1) -> both components of fc are estimated (DEFAULT).
Principal point optimized (center_optim=1) - (DEFAULT). To reject principal point, set center_optim=0
Skew not optimized (est_alpha=0) - (DEFAULT)
Distortion not fully estimated (defined by the variable est_dist):
Sixth order distortion not estimated (est_dist(5)=0) - (DEFAULT) .
Main calibration optimization procedure - Number of images: 20
Gradient descent iterations: 1...2...3...4...5...6...7...8...9...10...11...12...13...14...15...16...17...18...19...20...21...22...23...24...25...26...27...28...29...30...done
Estimation of uncertainties...done
Calibration results after optimization (with uncertainties):
Focal Length: fc = [ 997.53863 1001.94304 ] +/- [ 54.94370 55.38183 ]
Principal point: cc = [ 298.03007 249.25204 ] +/- [ 13.51600 7.55436 ]
Skew: alpha_c = [ 0.00000 ] +/- [ 0.00000 ] => angle of pixel axes = 90.00000 +/- 0.00000 degrees
Distortion: kc = [ 0.01788 1.17241 0.00353 -0.00713 0.00000 ] +/- [ 0.04494 0.74699 0.00278 0.00345 0.00000 ]
Pixel error: err = [ 0.18501 0.15435 ]
Note: The numerical errors are approximately three times the standard deviations (for reference).
Recommendation: Some distortion coefficients are found equal to zero (within their uncertainties).
To reject them from the optimization set est_dist=[0;1;1;1;0] and run Calibration
visualize_distortions
现在，就像练习(在实践中没有真正建议的那样)，让我们在没有镜头扭曲模型的情况下进行优化(通过强制kc =[0;0;0;0;0;0])并且强制fc为0，为此，在matlab提示符中设置二元变量est_dist为[0;0;0;0]和est_aspect_ratio为0:
>> est_dist = [0;0;0;0;0];
>> est_aspect_ratio = 0;
Aspect ratio not optimized (est_aspect_ratio = 0) -> fc(1)=fc(2). Set est_aspect_ratio to 1 for estimating aspect ratio.
Principal point optimized (center_optim=1) - (DEFAULT). To reject principal point, set center_optim=0
Skew not optimized (est_alpha=0) - (DEFAULT)
Distortion not fully estimated (defined by the variable est_dist):
Second order distortion not estimated (est_dist(1)=0).
Fourth order distortion not estimated (est_dist(2)=0).
Sixth order distortion not estimated (est_dist(5)=0) - (DEFAULT) .
Tangential distortion not estimated (est_dist(3:4)~=[1;1]).
Main calibration optimization procedure - Number of images: 20
Gradient descent iterations: 1...2...3...4...5...6...7...8...9...10...11...12...13...14...15...16...17...18...19...20...21...22...23...24...25...done
Estimation of uncertainties...done
Calibration results after optimization (with uncertainties):
Focal Length: fc = [ 859.95431 859.95431 ] +/- [ 84.85733 84.85733 ]
Principal point: cc = [ 362.68078 225.03187 ] +/- [ 12.32947 5.63310 ]
Skew: alpha_c = [ 0.00000 ] +/- [ 0.00000 ] => angle of pixel axes = 90.00000 +/- 0.00000 degrees
Distortion: kc = [ 0.00000 0.00000 0.00000 -0.00000 0.00000 ] +/- [ 0.00000 0.00000 0.00000 0.00000 0.00000 ]
Pixel error: err = [ 0.21106 0.17413 ]
Note: The numerical errors are approximately three times the standard deviations (for reference).
>> est_aspect_ratio = 0;
>> center_option = 0;
>> center_optim = 0;
>> est_dist = [0;0;0;0;0];
>> eat_alpha = 0;
>> cc = [(nx - 1)/2; (ny - 1)/2];
Aspect ratio not optimized (est_aspect_ratio = 0) -> fc(1)=fc(2). Set est_aspect_ratio to 1 for estimating aspect ratio.
Principal point not optimized (center_optim=0). Note: to set it in the middle of the image, clear variable cc, and run calibration again.
Skew not optimized (est_alpha=0) - (DEFAULT)
Distortion not fully estimated (defined by the variable est_dist):
Second order distortion not estimated (est_dist(1)=0).
Fourth order distortion not estimated (est_dist(2)=0).
Sixth order distortion not estimated (est_dist(5)=0) - (DEFAULT) .
Tangential distortion not estimated (est_dist(3:4)~=[1;1]).
Main calibration optimization procedure - Number of images: 20
Gradient descent iterations: 1...2...3...4...5...6...7...8...9...10...11...12...13...14...15...16...17...18...19...20...21...22...23...done
Estimation of uncertainties...done
Calibration results after optimization (with uncertainties):
Focal Length: fc = [ 1108.24978 1108.24978 ] +/- [ 27.98441 27.98441 ]
Principal point: cc = [ 319.50000 239.50000 ] +/- [ 0.00000 0.00000 ]
Skew: alpha_c = [ 0.00000 ] +/- [ 0.00000 ] => angle of pixel axes = 90.00000 +/- 0.00000 degrees
Distortion: kc = [ 0.00000 0.00000 0.00000 -0.00000 0.00000 ] +/- [ 0.00000 0.00000 0.00000 0.00000 0.00000 ]
Pixel error: err = [ 0.22425 0.19920 ]
Note: The numerical errors are approximately three times the standard deviations (for reference).
获取得到：
Calib_Results_left.mat
Right标定同上：获取得到：
Calib_Results_right.mat
3. MATLAB之双目标定
>> stereo_gui
Loading of the individual left and right camera calibration files
Name of the left camera calibration file ([]=Calib_Results_left.mat): Calib_Results_left.mat
Name of the right camera calibration file ([]=Calib_Results_right.mat): Calib_Results_right.mat
Loading the left camera calibration result file Calib_Results_left.mat...
Loading the right camera calibration result file Calib_Results_right.mat...
Stereo calibration parameters after loading the individual calibration files:
Intrinsic parameters of left camera:
这样整个标定就完成了。。。。。。。。。。。