使用MATLAB的方法参考:
Matlab单目相机+投影仪标定(cam_proj_calib)
1.结构光相机标定软件
结构光相机标定是一件十分复杂的过程,所幸布朗大学开源了一套投影仪摄像机标定系统(Projector-Camera Calibration/3D Scanning Software)
参考网址:http://mesh.brown.edu/calibration/
http://mesh.brown.edu/calibration/software.html
官网提供了Window和Apple OSX编译好的标定软件,版本比较老旧,支持Windows 7 x64,Opencv也是比较老旧的OpenCV 2.4.11
需要安装VS2010和QT
参考:Visual Studio 2010 部署QT4.8.4
运行程序,出现如下界面
选择图像文件夹
选择calib-small
获取图像
修改参数
选择校准
校准流程
标定完成
标定结果如下:
Camera:
- reprojection error: 0.210457
- K:
[2472.409411080203, 0, 1080.85167590167;
0, 2478.017886117078, 593.2818615670036;
0, 0, 1]
- kc: [-0.06941443683216893, -0.4695970130189265, -0.01160410430692993, 0.001515755788593767, 0]
Projector:
- reprojection error: 0.22941
- K:
[2348.989003513338, 0, 543.4519374462692;
0, 2337.279628591229, 829.6031628250041;
0, 0, 1]
- kc: [-0.1053159744264396, 0.3851158289564819, -0.01584415581400273, -0.002263428058331434, 0]
Stereo:
- reprojection error: 0.440135
- R:
[0.92305185328739, -0.01740931313743194, -0.3842813968419462;
0.04436690718206724, 0.9971268756188638, 0.06139683595805918;
0.3821084318486211, -0.07372184028255018, 0.921172207882729]
- T: [345.4749488391364;
-264.2733060739041;
199.1634577672953]
标定结果
解码的投影仪列和行图像