Recognition and Pose Estimation of 3D Objects through 3D Features

by Tolga Birdal for Open Source Computer Vision Library (OpenCV)

Enabling robots to automatically locate and pick up randomly placed objects from a bin is crucial in factory automation replacing tedious and heavy manual labor. Similar tools are also capable of guiding impaired people/robots through unstructured environments (automated navigation). These make 3D matching a ubiquitous necessity. Within the context of this application, I will be proposing an OpenCV implementation of a 3D object recognition and pose estimation algorithm using 3D point features.