AVILUSplus

AVILUSplusLogo

Project Manager: Prof. Dr. Didier Stricker
Funding by: BMBF

AVILUSplus is a research and development project funded by the Federal Ministry of Education and Research (BMBF). It is in close relationship with the German technology network AVILUS, where participants of the German core industries and applied research develop innovative technologies in the areas of virtual (VR) and augmented reality (AR) for product and resource life cycle.

The AVILUSplus consortium consists of nine research institutions, one of which is the German Center for Artificial Intelligence (DFKI GmbH Kaiserslautern). AVILUSplus focuses on long-term technologies with needs for fundamental research. The main areas are: information management, simulation and rendering, tracking, interaction and geometry acquisition.

The research department “Augmented Vision” participates in the work packages tracking and interaction with application to AR-assisted industrial maintenance (“AR Handbook”). The goal is to develop innovative methods for image-based tracking and interaction with help of spherical cameras. A head-worn spherical camera provides a field of view that is similar to the human visual system. It captures not only a central part of the frontal working area, as a standard perspective camera would do, but also the hands of the user and parts of the environment. Moreover, the change in the image content is comparatively small, even under large camera motions. For these reasons, spherical cameras can be used to solve image-based tracking and interaction with only one sensor.

Real-time camera tracking with spherical cameras

The work of the department “Augmented Vision” focuses on the following aspects with special respect to the geometry of spherical cameras: modelling and automatic calibration, landmark detection, description and tracking, pose estimation and error propagation, object recognition, hand detection and gesture recognition. First results include the automatic calibration of spherical cameras with a simple checkerboard pattern, real-time tracking with circular markers, and color-, contour- and shading-based hand detection.

For further details about the project, please visit the project webpage: http://www.avilusplus.de.