c – OpenCV – 从视差图计算实际距离

这似乎是一个相对容易的问题.根据我的理解,我可以简单地使用公式

distance = (baseline * focal length) / disparity

我可以使用矩阵Q(stereoRectify输出). Q [2] [3] =焦距,1 / Q [3] [2] =基线.

计算出的Q矩阵是：

Q: !!opencv-matrix
   rows: 4
   cols: 4
   dt: d
   data: [ 
    1.,0.,-1.5668458938598633e+02,1.,-1.1948609733581543e+02,2.3598119491957863e+02,1.6254073321947445e-02,0. ]

问题是结果与现实不符.例如,对于瞄准房间天花板的相机,距离约为. 2.5米(其中视差被正确地计算为12),实际距离计算为1,3米.对于非常近的物体(例如30厘米),它似乎是正确的,但远物体是非常不正确的.在校准过程中,我指定了棋盘方块的确切尺寸,单位为毫米.

我做了以下几点：

// compute rectification transforms from calibration data
stereoRectify(M1,D1,M2,D2,Size(FRAME_WIDTH,FRAME_HEIGHT),R,T,R1,R2,P1,P2,Q,CALIB_ZERO_DISPARITY,0.0,&roi1,&roi2);

// compute the undistortion and rectification transformation maps for each camera
initUndistortRectifyMap(M1,CV_16SC2,map11,map12);
initUndistortRectifyMap(M2,map21,map22);

...

// get images from camera (VideoCapture)
camLeft.read(_frameLeft);
camRight.read(_frameRight);

// remap images using the calibration data
remap(_frameLeft,frameLeft,map12,INTER_LINEAR);
remap(_frameRight,frameRight,map22,INTER_LINEAR);

// compute disparity from undistorted images
stereo->compute(frameLeft,disparityMap);

...

// compute the real-world distance [mm]
float fMaxDistance = static_cast<float>((1. / Q.at<double>(3,2)) * Q.at<double>(2,3));

// outputDisparityValue is single 16-bit value from disparityMap
// DISP_SCALE = 16
float fDisparity = outputDisparityValue / (float)StereoMatcher::DISP_SCALE;
float fDistance = fMaxDistance / fDisparity;

我做错了吗？
提前致谢.