This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
This page shows OpenCV functionalities for stereo. See also the sample in cpp stereo_match.cpp or python2/stereo_match.py (more simple and readable to start, it also creates a 3D mesh viewable by MeshLab).
These examples compute the disparity map. To track your interesting points, you need to perform tracking on both images (right and left), rectify images and find points association (using epipolar geometry). After you need the "original" 3D position of points to compute the calibration step.
If you want to calibrate your stereo system, you probably should use a chessboard pattern and use OpenCV function here to calibrate and findChessbordCorner for a chessboard or findCircleGrid. See samples in cpp/tutorial_code/calib3d and the related pages on wiki.
If you want to track your interesting points in 3D. Perform the calibration step, compute your points in both images and use the association and disparity map to get the 3D position estimation.
