The concept of Harvest4D is based on the observation that there is a strongly growing and almost ubiquitous availability of sensing devices delivering different kinds of data streams. The research team of Harvest4D therefore proposes a radical paradigm change for acquisition and processing of visual 3D data to exploit this observation. New ways of capturing, processing and displaying visual 3D data are developed to move away from the classical planning and goal-oriented acquisition methodology determining sensor devices and acquisition data in advance. In contrast to this classical approach, the Harvest4D paradigm change enables sensor devices and available data themselves defining the process.

The goal of this methodology is to enable easy incidental data capturing in addition to people's actually deliberate tasks while prevailing reconstruction of high quality 3D visual models.
Different challenging problems arise from this new concept, such as dealing with continuous streams of time-dependent data, finding means of integrating data from different sensors and modalities, detecting changes in data sets for 4D models and many more.

See the official project page for more information:

General Information

Project Leader:
Michael Wimmer

Participating Research Staff:
Michael Goesele
Samir Aroudj

EU 7th Framework Program

Project duration:
June 2013 – May 2016