Module 10: Camera Calibration & 3D Vision

Camera calibration, stereo vision, 3D reconstruction, and Structure from Motion.

Topics Covered

  • Camera calibration and undistortion
  • Pose estimation with PnP
  • Stereo vision and disparity maps
  • 3D reconstruction and point clouds
  • Structure from Motion (SfM)

Tutorial Files

File Description
01_camera_calibration.py Calibration, undistortion, perspective transform
02_pose_estimation.py 3D pose estimation with PnP
03_stereo_vision.py Stereo calibration, rectification, disparity maps
04_3d_reconstruction.py Depth estimation and point cloud generation
05_sfm_concepts.py Structure from Motion fundamentals

Key Concepts

Camera Matrix K 

K = [fx  0  cx]
    [0  fy  cy]
    [0   0   1]

Stereo Pipeline 

Calibration → Rectification → Matching → Disparity → Depth

Key Functions Reference

Function Description
cv2.findChessboardCorners() Detect calibration pattern
cv2.calibrateCamera() Calibrate camera
cv2.undistort() Remove lens distortion
cv2.solvePnP() Estimate 3D pose
cv2.stereoCalibrate() Calibrate stereo pair
cv2.stereoRectify() Rectify stereo images
cv2.StereoBM_create() Block matching disparity
cv2.StereoSGBM_create() Semi-global matching

Further Reading