Multi-view Continuous Structured Light Scanning

Fabian Groh, Benjamin Resch and Hendrik P. A. Lensch

University of Tübingen

For objects (a) with optically challenging properties, shifted sinusoidal patterns (b) are captured from multiple cameras and projectors. On top of the phase information, highly accurate features (c) are generated to perform a precise bundle adjustment (d). With the multi-view continuous phase signals (e) the reliable camera-projector pairs for every surface point are estimated by validating multi-view consistency (f). Finally, we optimize the depth over all available information.

We introduce a highly accurate and precise multi-view, multi-projector, and multi-pattern phase scanning method for shape acquisition that is able to handle occlusions and optically challenging materials. The 3D reconstruction is formulated as a two-step process which first estimates reliable measurement samples and then simultaneously optimizes over all cameras, projectors, and patterns. This holistic approach results in significant quality improvements. Furthermore, the acquisition time is drastically reduced by relying on just six high-frequency sinusoidal captures without the need of phase unwrapping, which is implicitly provided by the multi-view geometry.

GCPR 2017

Paper

Video Mug/Angel

Bibtex