Skip to Main Content
A system for fast multi-view 3D model reconstruction of object sequences is composed of a number of hardware and software components: the multiple simultaneous image acquisition subsystems, the computation platform, the object/background segmentation algorithm, and in this case, a volumetric carving procedure based on the silhouettes of the objects from each view that generates a volumetric representation of the object under analysis. This volumetric representation can be then transformed into a 3D model applying isosurface generation techniques. However, as it will be shown, before any of these steps is performed a finely tuned camera calibration has to be obtained. The accuracy of this calibration is essential for the quality of the generated model. This paper describes the structure of a fast 3D reconstruction system, analyzes its features and focuses on the new techniques developed to enhance the calibration of its multiple cameras.