Course materials for MIPT Camera Geometry for Scene Reconstruction course (Cognitive Technologies)
Basic principles of pinhole camera projection and Homographies are illustrated. The final task is to project image points from lidar to camera on example from KITTI dataset
Camera calibration with a chessboard pattern is performed, using opencv gradients calculation as well as numerical derivatives. Optimisation is performed via scipy.optimize
This is an illustration of how depth map is calculated using opencv SGBM functions. It doesn't contain programming taks, but requires answering some theoretical questions.