Computational Imaging for Scene Understanding

3Time-of-Flight and Transient Rendering

Adithya Kumar PEDIREDLA

Dartmouth College, Hanover, USA

3.1. Introduction

Time-of-flight rendering (Jarabo et al. 2014; Marco et al. 2019; Pediredla et al. 2019b, 2020) refers to generating an image or sequence of images that a time-of-flight camera captures for a given scene. Typically, a time-of-flight render, capable of time-of-flight rendering, is realized by replacing intensity cameras in a physics-based rendering engine with time-of-flight cameras. There are several time-of-flight cameras on the market, such as the Swiss ranger (Cazorla et al. 2010), Kinect (Smisek et al. 2013), intensified charged coupled device (ICCD) (Cester et al. 2019), single-photon avalanche diode detector (Villa et al. 2014) and Hamamatsu streak cameras. They are employed for depth sensing, fluorescence lifetime imaging, transient imaging, non-line-of-sight imaging and imaging through a scattering medium. At an abstract high level, the time-of-flight measurements are equal to the weighted sum of photon intensities with a weight (|w| ≤ 1) that depends on the photon’s time-of-travel¹ and the design of the time-of-flight camera. More details on this weighting function are in section 3.2.1. A time-of-flight renderer simulates a time-of-flight camera by implementing the weighting function accurately and for scenes with arbitrary scene geometry, material parameters and camera optics.

Time-of-flight rendering finds several applications in sensor design, machine ...