Research Projects (1996)

Uncalibrated stereo, or more precisely the recovery of the epipolar geometry and the projective structure of a real world object purely from the knowledge of homologue points, is an active field of research. The application is mainly found in robot vision as one has to deal with a permanently changing image acquiring situation not allowing an absolute calibration. The aim of the project is to apply this technique for stationary cameras for which an absolute calibration is hard to perform or even impossible (for example a scanning electron microscope). The core of the system is the integration of physics based vision and shape from stereo in the uncalibrated case. For this purpose robust shape from stereo techniques are developed. The physics based vision algorithms are especially adapted to the situation in the scanning electron microscope. The project is strongly related to the project "Application of Image Processing in Materials Science"

Interactive multimedia, based on computer networks, may be of great benefit for many disciplines. That is in specific true for remote sensing (earth observation), where daily a large amount of image-data is collected over the entire globe. By multimedia, the most recent satellite data and associated meta-information is available to every potential user.

Within this project, key services and tools are implemented, in order to keep the Austrian user community up-to-date and to support them on the fast developing international market.

Another goal is performed in cooperation with the "Environmental Planning" division at Austrian Research Centers / Seibersdorf: By merging multisensoral (radar/optical) and multitemporal data, new methods for monitoring agricultural sites are being developed.

The goal of the VIRGO network is to coordinate European research and postgraduate training activities that address the development of intelligent robotic systems able to navigate in unknown and possibly dynamic environments. This goal will be achieved through a framework which enhances RTD activities in European laboratories, already established in the aforementioned scientific area. Specifically alternative environment representations based on visual information will be studied and methods to process these representations and use them to control the motion of a robot will be developed. Environment learning issues will also be considered, aiming at autonomous acquisition of the discriminating environment features. The task of ICG in the network is to coordinate research in Fusion.

Partnerlist: List of partners maintained by project coordinator

The series production of parts with a small to medium number of units requires a fast and inexpensive adaption to new parts. A pure hardware solution, as it is common for example in the motor industry, is contrary to the above requirement. The goal of this project is to develop a vision driven, flexible and automatic assembling unit. In order to solve this demanding task, the separation into three, from the viewpoint of vision almost independent steps is proposed. The first step is to isolate one part from the pile. This task is generally referred as binpicking. Where the task of step one was to detect one plane among many parts of one class the task now is to determine the exact position of one known part (step 2). And the third step is the surveillance of the assembling itself.