Computational Imaging for Scene Understanding

8Synthetic Wavelength Imaging: Utilizing Spectral Correlations for High-Precision Time-of-Flight Sensing

Florian WILLOMITZER

Wyant College of Optical Sciences, University of Arizona, Tucson, USA

8.1. Introduction

Optical three-dimensional (3D) imaging and ranging techniques have been used in academia and industry for many years with great success. Fields of application include medical imaging, autonomous navigation, industrial inspection, forensics or virtual reality. The great success of these techniques is no coincidence, as high-quality 3D object or scene representations contain a distinctive higher information content than simple 2D images: compared to 2D images, 3D representations are invariant against translation and rotation of the object as well as variations in surface texture or external illumination conditions. This fact has made 3D imaging an established tool in optical metrology and, more recently, in computer vision.

Although the sheer number of available 3D sensing principles is immense, existing approaches can be broadly categorized into three groups: (1) triangulation-based approaches (including active stereo, passive stereo or focus search) (Takeda and Mutoh 1983; Srinivasan et al. 1984; Schaffer et al. 2010; Willomitzer et al. 2010; Ettl et al. 2013; Arold et al. 2014; Schönberger and Frahm 2016; Willomitzer and Häusler 2017), (2) reflectance-based approaches that measure the surface gradient (including photometric stereo and deflectometry) (Woodham 1980 ...