The projects idea was to create an automatic object detection (of different colored toy like containers) and process their 3D position and rotation in different light conditions so that a raspberry pi controlled container roboter can pick them up and position them with a different script. the project hast two different implementations for the Pose Estimation, both use opencv, no deep learning. One is a single camera single shot solution using opencvs solvePnP function. The other is a stereo camera solution. Both can use charuco or aruco boards for camera calibration. Preprocessing includes color quantization, camera calibration, Edge Detection, HSV Masks, Morphology Operations, curve detection.