Cognitive 3D Image Processing for Robotic Applications

Current approaches in the field of image processing rely heavily on local geometries. This knowledge can be applied to intelligent image processing, especially in industrial scenarios where scenes are highly structured with respect to functionality and geometries.

PROFACTOR develops algorithms that build a bridge between traditional low-level feature methods and high-level semantic concepts. Knowledge of the industrial problem to be tackled is thus mapped by semantic concepts. The results are novel image processing systems that can be commissioned and operated by non-experts. These systems cover a wide spectrum of applications from classical 3D positioning and identification, through quality testing, to identification and verification of unknown production workspaces.