3D-Inside

Reconstruction of large infrastructure interiours like railway stations in order to provide a useable 3D model with important other information like pedestrian simulation, viewlines, exchangeable storefronts and advertisements etc.

The management of edifices and buildings of large public transport infrastructures requires up to date information about the current status of a building, the way it is used by people and how certain modifications would affect the whole infrastructure management. Interactive navigation, simulation and visualization in three dimensions can greatly facilitate providing such information. The overall objective of 3D-INSIDE is to combine novel techniques of visual computing into a prototype for the 3D visualization, navigation and pedestrian simulation within the indoor environment of public transport infrastructures. The consortium of 3D-INSIDE proposes to (1) develop techniques for automatically and rapidly obtaining 3D models of the interior of an infrastructure; (2) develop realistic simulations of not yet well-understood pedestrian walking behaviour within transport infrastructures, by incorporating real video-based pedestrian data obtained by advanced algorithms for vision-based people detection and tracking; (3) integrate both the results from vision-based modelling and pedestrian simulation into an efficient 3D viewer application. 3D-INSIDE will enable to establish several links between visual computing disciplines. The image-based 3D reconstruction of an indoor environment will allow determining a priori camera positions and fields of view for measuring pedestrian trajectories, and will allow incorporating context knowledge into pedestrian detection, resulting in reduced computational complexity and more reliable trajectories.

Until the year 2011 more than 10 major Austrian railway stations will be remodelled or newly built, moving investments of over a billion Euro in total. 70% of the industrial partners' (Austrian Federal Railways) revenues are made at the 28 most frequented train stations. An easily updated 3D overview of the buildings and a better understanding of the motion behaviour of the passengers will simplify the maintenance and management of these stations and make it possible to offer improved services to the more than 360.000 customers daily. To understand and model the behaviour of 45% of the daily clients and to be able to visualize it for business partners and infrastructure managers alike will be an important marketing tool. The methods developed within this project can also be sold to other European infrastructure operators, airports, and shopping centres.