双目视觉嵌入式移植后校正矩阵的生成方法

最新推荐文章于 2019-05-07 08:50:49 发布

星之河

最新推荐文章于 2019-05-07 08:50:49 发布

阅读量1.5k

点赞数

分类专栏：嵌入式Linux 双目立体视觉算法与程序实现文章标签： linux opencv QT 校正双目视觉

本文链接：https://blog.csdn.net/b5w2p0/article/details/39099177

版权

嵌入式Linux 同时被 2 个专栏收录

76 篇文章 1 订阅

订阅专栏

双目立体视觉算法与程序实现

10 篇文章 51 订阅

订阅专栏

首先要利用标定得到的参数，通过调用OpenCV中的函数来得到校正矩阵。

   
    double LM[] = {

     4.0178810502329486e+002, 0, 1.5950000000000000e+002,
     0,4.0178810502329486e+002,1.1950000000000000e+002,
     0,0, 1
     };
     double LD[] = {

    -4.5141575299933429e-001, 2.3998422552817714e-001,
     1.5081252615814692e-002, 1.8419337906291466e-003, 0
     };
     double RM[] = {

     4.0178810502329486e+002, 0, 1.5950000000000000e+002,0,
     4.0178810502329486e+002, 1.1950000000000000e+002,0,0, 1
     };
     double RD[] = {

     -5.1762688060193929e-001,6.3474560985437023e-001,
     -2.5303249541511292e-003, -2.7165709692779268e-003, 0
     };
     double Rotate[] = {

     9.9952858710662951e-001,1.4994693512808868e-002,
     2.6791094100126738e-002,-1.1568939379481839e-002,
     9.9224798371307799e-001,-1.2373398263599662e-001,
     -2.8438762249064085e-002, 1.2336570829767958e-001,
     9.9195369792039845e-001
     };
     double translation[] = {

     -8.8108837791674333e+001, -1.3412948452799096e+000,
     9.4102431499505457e+000
     };



cv::Mat    cameraMatrix_L = cv::Mat(3,3,CV_64FC1,LM);
    cv::Mat	distortionCoefficients_L = cv::Mat(1,5,CV_64FC1,LD);
    cv::Mat    cameraMatrix_R = cv::Mat(3,3,CV_64FC1,RM);
    cv::Mat	distortionCoefficients_R = cv::Mat(1,5,CV_64FC1,RD);

    cv::Mat rotations= cv::Mat(3,3,CV_64FC1,Rotate);
    cv::Mat translations= cv::Mat(3,1,CV_64FC1,translation);
    cv::Mat R1, R2, P1, P2, Q;
    cv::Rect roi1,roi2;
    cv::Mat m_Calib_Mat_Remap_X_L  = cv::Mat(imageSize, CV_32FC1);
    cv::Mat m_Calib_Mat_Remap_Y_L= cv::Mat(imageSize, CV_32FC1);
    cv::Mat m_Calib_Mat_Remap_X_R = cv::Mat(imageSize, CV_32FC1);
    cv::Mat m_Calib_Mat_Remap_Y_R = cv::Mat(imageSize, CV_32FC1);

    cv::stereoRectify(
         cameraMatrix_L,
         distortionCoefficients_L,
         cameraMatrix_R,
         distortionCoefficients_R,
         imageSize,
         rotations,
         translations,
         R1,R2,P1,P2,Q,
         1024
         );
    initUndistortRectifyMap(
         cameraMatrix_L,
         distortionCoefficients_L,
         R1,P1,
         imageSize,
         CV_16SC2,
         m_Calib_Mat_Remap_X_L, m_Calib_Mat_Remap_Y_L);
     initUndistortRectifyMap(
         cameraMatrix_R,
         distortionCoefficients_R,
         R2, P2,
         imageSize,
         CV_16SC2,
         m_Calib_Mat_Remap_X_R,m_Calib_Mat_Remap_Y_R);

嵌入式运行后可以得到相关校正参数。