Intelligent Systems Research Group

The focus of the project is on the research of intelligent image based 3D sensor systems. Besides the delivery of 3D data, the sensors will incorporate image processing algorithms to provide to the user additional information (like for example deviations from a prescribed geometry of the objects under investigation).

For this, a modular and service oriented system architecture will be foreseen. A semantic user interface to the system will ease its use in industry 4.0 application. Exemplarily for the integration of hardware and software in the system architecture, two types of 3D sensors will be integrated: deflectometric sensors (active operation) and plenoptic sensors (passive operation). Characteristic for these two sensors are the complexity of the signal- and image processing algorithms. These require both high processing power as well as big efforts and expert knowledge to adapt them to a specific application.

Due to the demanding algorithms, FPGAs will be used to implement the 3D image processing algorithms. Adaptation of the algorithms to pipeline operation and parallel processing guarantees fast availability of the results. By using technical and physical sensor models, as well as models for ambient impairments and a context model for describing the use case, intelligent configuration, calibration, parameterization and adaptation to changing environment will be possible.

The result of the project is a cyber-physical system which offers its services for optical acquisition and processing of 3D data.  A prototype of the system will be implemented and evaluated for typical industrial tasks.