Feature-Based Object Detection and Pose Estimation Based on 3D Cameras and CAD Models for Industrial Robot Applications

This paper presents a feature-based object detection and pose estimation method. In this approach, a user selects geometric features from a CAD model of an object. The selected features are then matched against measured features from the 3D cameras. Software modules were developed for the method and were tested in a robot cell. Based on the results, our approach provides a fast way to configure and program the pose estimation system for new objects. Target applications of the approach are in small series and agile, even one-of-a-kind manufacturing.