Researching new methods and systems for automated visual inspection

The Variable Image Acquisition and Processing research group develops methods and systems for automated visual inspection that exploit various kinds of heterogeneous information.

Research areas and competences

We study theoretical and practical issues related to variable image acquisition, i.e. how choices are made during inspections when a single image does not fully capture the features of interest. Our focus areas include modeling image acquisition based on holistic systems theory, the reproducible acquisition of optimal image series, and controlling the variable acquisition parameters – such as illumination, focus, camera position and spectral sensitivity (active vision)
– online. Our expertise also extends to methods like ellipsometry and deflectometry. Combined with our capabilities in data fusion, our solutions yield a description quality that is second to none.

Applications and products

We tackle cases that typically cause problems for automated visual inspection systems, such as partially or fully specular (i.e. mirroring) surfaces, transparent objects and thin films on curved surfaces. Our particular skills set can also help solve issues in other fields, like classifying moving objects, providing automated interactive assistance during surgery and generating open, adaptive models of the environment for artificial cognitive systems.

The VBV research group does not belong to any of the scientific departments; instead, it forms the institutional link between Fraunhofer IOSB and Karlsruhe Institute of Technology’s Vision and Fusion Laboratory, the chair held by Fraunhofer IOSB head of institute Prof. Beyerer.