Optimizing Range and Velocity Sensing with Computational Single-photon Imaging
Abstract
The capacity of imaging systems must continue to expand to keep pace with rapidly emerging technologies. Currently, imaging systems are not capable of quickly capturing 3D geometry and object velocity simultaneously, nor are they able to efficiently resolve non-line-of-sight scenarios such as occluded scene parts or objects veiled by scattering. These deficiencies impact their usefulness in mapping terrain or for providing reliable sensing for navigation of autonomous vehicles. Optimized computational range and velocity imaging provides a solution for the above-mentioned challenges and facilitates entirely new modalities in other applications. For example, imagine a self-driving vehicle equipped with sensing technology that allows for range and motion of its environment to be captured instantaneously at high resolution and high speed. Further, envision the same sensing mechanism to be able to clearly see through fog, snow, and smoke (scattering media) or to Òsee around objectsÓ (non-line-of-sight imaging) to detect what lies beyond the next tree or hill. Such a sensing technology would help make self-driving cars safe and advance the sensing capabilities of other types of autonomous vehicles. Computational imaging places these and other unprecedented modalities within reach. The goal of this project is to transcend the boundaries of existing imaging and remote sensing systems by optimizing the speed, resolution, precision of depth and velocity estimation, and 3D imaging capabilities of emerging computational single-photon imaging systems for direct and non-line-of-sight scenarios.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 11, 2019
- Source ID
- W911NF1910120
Entities
People
- Gordon Wetzstein
Organizations
- Army Contracting Command
- Stanford University
- United States Army