Advanced Unsupervised Monitoring and Visualization of Complex Scenarios
Project partners: CGV, ICG, Arsenal, Siemens. This project was funded by FIT-IT.
Large areas with CCTV surveillance, such as train stations or shopping malls, have a high number of cameras. On a typical montior array it is not easy to get an overview or to follow a person through multiple camera images. In contrast, in our project we use an integrated 3D model of the entire site. Automatically detected persons are shown as a billboard in the 3D scene and camera images are backprojected into the model.
The building was modeled using the GML (Generative Modeling Language)
. Several levels of detail are used for visualization. At the finest level, video images from the CCTV are backprojected in the 3D model. For detected persons, billboards are inserted at the respective location. In contrast to a traditional monitor array, in our system the operator can easily follow a person that walks through several camera images.
The Graz HEyeWall is a 4m by 2m rear projection display with 8 Megapixels, ideal for such a large amount of data to be displayed simultaneously.