Photogrammetry based systems are able to produce 3D reconstructions of an object given a set of images taken from different orientations. In this paper, we implement a light-field camera within a photogrammetry system in order to capture additional depth information, as well as the photogrammetric point cloud. Compared to a traditional camera that only captures the intensity of the incident light, a light-field camera also provides angular information for each pixel. In principle, this additional information allows 2D images to be reconstructed at a given focal plane, and hence a depth map can be computed. Through the fusion of light-field and photogrammetric data, we show that it is possible to improve the measurement uncertainty of a millimetre scale 3D object, compared to that from the individual systems. By imaging a series of test artefacts from various positions, individual point clouds were produced from depth-map information and triangulation of corresponding features between images. Using both measurements, data fusion methods were implemented in order to provide a single point cloud with reduced measurement uncertainty.

Date Published: 23 August 2017
PDF: 8 pages
Proc. SPIE 10373, Applied Optical Metrology II, 1037304 (23 August 2017); doi: 10.1117/12.2272999

Published in SPIE Proceedings Vol. 10373:
Applied Optical Metrology II
Erik Novak; James D. Trolinger, Editor(s)