I done all step for stereo calibration but I getting a just a white screen then I tried your stereo calibration file it's working it showing some depth but it's not going to be correct depth because I am using different from your camera (two c270 camera).

this is my stereo file
%YAML:1.0

K1: !!opencv-matrix
rows: 3
cols: 3
dt: d
data: [ 8.9097802803323373e+02, 0., 3.0006642964743565e+02, 0.,
9.2593842191950148e+02, 7.8165024820505010e+01, 0., 0., 1. ]
D1: !!opencv-matrix
rows: 1
cols: 5
dt: d
data: [ 4.5571307763078717e-01, -2.1371653403804274e+00,
-1.0600088198594815e-01, -4.1709276363259854e-03,
7.6138794069241875e+00 ]
K2: !!opencv-matrix
rows: 3
cols: 3
dt: d
data: [ 8.8097737167946809e+02, 0., 3.3341133238820430e+02, 0.,
9.1397906735611662e+02, 1.0851476284668595e+02, 0., 0., 1. ]
D2: !!opencv-matrix
rows: 1
cols: 5
dt: d
data: [ 9.2906895358031608e-02, 1.0129152698060886e+00,
-9.5093716749191579e-02, 4.5423297986634187e-03,
-2.3420055234373098e+00 ]
R: !!opencv-matrix
rows: 3
cols: 3
dt: d
data: [ 9.9981362129215290e-01, -1.7125721991468577e-02,
-8.9124814132944419e-03, 1.6998282631813209e-02,
9.9975491330655863e-01, -1.4183501223761859e-02,
9.1531997815183005e-03, 1.4029360843117065e-02,
9.9985968814033765e-01 ]
T: !!opencv-matrix
rows: 3
cols: 1
dt: d
data: [ 1.0577134040602316e-01, 1.4328375860577982e-03,
-1.1656850572458893e-02 ]
E: !!opencv-matrix
rows: 3
cols: 3
dt: d
data: [ 2.1126148930712530e-04, 1.1674095429020535e-02,
1.2673015875918243e-03, -1.2622824193606941e-02,
-1.2842723192162449e-03, -1.0565260796848723e-01,
3.6536060292735226e-04, 1.0576995563609995e-01,
-1.4874377977337520e-03 ]
F: !!opencv-matrix
rows: 3
cols: 3
dt: d
data: [ 1.9124993696215748e-07, 1.0169253916597929e-05,
1.6991364161982508e-04, -1.1014530791674363e-05,
-1.0783294803252525e-06, -7.8750996918197716e-02,
1.4228594622010579e-03, 7.7895980868225079e-02, 1. ]
R1: !!opencv-matrix
rows: 3
cols: 3
dt: d
data: [ -9.9293279311101512e-01, 5.0973054514407415e-03,
1.1856848587163883e-01, -4.2910791113344423e-03,
-9.9996591349316033e-01, 7.0539699354129530e-03,
1.1860040052554023e-01, 6.4953313175043813e-03,
9.9292082044151808e-01 ]
R2: !!opencv-matrix
rows: 3
cols: 3
dt: d
data: [ -9.9389176604265628e-01, -1.3463812346450965e-02,
1.0953484901943035e-01, 1.2771990513931762e-02,
-9.9989382612060951e-01, -7.0151795558210583e-03,
1.0961767034069771e-01, -5.5733511452202037e-03,
9.9395820038173144e-01 ]
P1: !!opencv-matrix
rows: 3
cols: 4
dt: d
data: [ 3.7013994770355175e+01, 0., 6.1154418468475342e+01, 0., 0.,
3.7013994770355175e+01, 5.5961920452117920e+02, 0., 0., 0., 1.,
0. ]
P2: !!opencv-matrix
rows: 3
cols: 4
dt: d
data: [ 3.7013994770355175e+01, 0., 6.1154418468475342e+01,
-3.9390806669324721e+00, 0., 3.7013994770355175e+01,
5.5961920452117920e+02, 0., 0., 0., 1., 0. ]
Q: !!opencv-matrix
rows: 4
cols: 4
dt: d
data: [ 1., 0., 0., -6.1154418468475342e+01, 0., 1., 0.,
-5.5961920452117920e+02, 0., 0., 0., 3.7013994770355175e+01, 0.,
0., 9.3966074574399450e+00, 0. ]

square_size 0.025 m 9 6 chessboard

what I missed

thanks

In stereo_depth.py 's Undistortion and Rectification part

leftMapX, leftMapY = cv2.initUndistortRectifyMap(K1, D1, R1, P1, (width, height), cv2.CV_32FC1)
left_rectified = cv2.remap(leftFrame, leftMapX, leftMapY, cv2.INTER_LINEAR, cv2.BORDER_CONSTANT)
rightMapX, rightMapY = cv2.initUndistortRectifyMap(K2, D2, R2, P2, (width, height), cv2.CV_32FC1)
right_rectified = cv2.remap(rightFrame, rightMapX, rightMapY, cv2.INTER_LINEAR, cv2.BORDER_CONSTANT)

I think it is not needed to initUndistortRectifyMap for every frame. It could have been done just for the first frame. is my understanding correct ?

-- Thank you,
Balaji

您好，这个实现起来误差有多大，是否做了空洞填充

Single camera calibration
python single_camera_calibration.py --image_dir data\leftFixedStereo --image_format png --prefix left --square_size 0.025 --width 9 --height 6 --save_file left_cam.yml
python single_camera_calibration.py --image_dir data\rightFixedStereo --image_format png --prefix right --square_size 0.025 --width 9 --height 6 --save_file right_cam.yml

Stereo camera calibration
python stereo_camera_calibration.py --left_file left_cam.yml --right_file right_cam.yml --left_prefix left --right_prefix right --left_dir data\bothImagesFixedStereo --right_dir data\bothImagesFixedStereo --image_format png --square_size 0.025 --save_file stereo_cam.yml

Disparity map
python stereo_depth.py --calibration_file stereo_cam.yml --left_source data\left_1541067450.avi --right_source data\right_1541067450.avi --is_real_time 0

use default data/* yml

Stereo camera calibration
python stereo_camera_calibration.py --left_file data\left_cam.yml --right_file data\right_cam.yml --left_prefix left --right_prefix right --left_dir data\bothImagesFixedStereo --right_dir data\bothImagesFixedStereo --image_format png --square_size 0.025 --save_file stereo_cam.yml

Disparity map
python stereo_depth.py --calibration_file data\stereo_cam.yml --left_source data\left_1541067450.avi --right_source data\right_1541067450.avi --is_real_time 0

In stereo_camera_calibration.py line: 103
image_size = gray_right.shape
I think it should be changed like this：
image_size = gray_right.shape[::-1]