Computational imaging with single-photon detectors
Abstract/Contents
- Abstract
- Active 3D imaging systems, such as LIDAR, are becoming increasingly prevalent for applications in human-computer interaction, robotics, autonomous vehicle navigation, remote sensing, and more. However, the ability to image with short acquisition times or at long distances is fundamentally limited by the weak signal of backscattered light. Moreover, conventional active imaging systems fail to exploit information contained in multiply-scattered light. In this thesis, we develop novel computational imaging frameworks for photon-efficient 3D imaging and imaging with multiply scattered light using time-resolved detectors that are sensitive to single photons. Our framework for photon-efficient 3D imaging uses convolutional neural networks to jointly process measurements from a single-photon detector and a conventional camera. The approach outperforms existing techniques in a challenging regime where the average number of detected signal photons at each pixel is less than one. Our frameworks for imaging with multiply scattered light efficiently model and invert light transport, enabling new capabilities in non-line-of-sight imaging (i.e., imaging around corners or behind occluders) and imaging through scattering media. For example, we demonstrate significant improvements to the computational efficiency and accuracy of non-line-of-sight imaging, enabling reconstruction of large hidden scenes with unprecedented resolution and fidelity. Our method for 3D imaging through scattering media efficiently models light propagation through a diffusive barrier and outperforms techniques like time- or coherence-gating.
Description
| Type of resource | text |
|---|---|
| Form | electronic resource; remote; computer; online resource |
| Extent | 1 online resource. |
| Place | California |
| Place | [Stanford, California] |
| Publisher | [Stanford University] |
| Copyright date | 2020; ©2020 |
| Publication date | 2020; 2020 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Author | Lindell, David Brian |
|---|---|
| Degree supervisor | Wetzstein, Gordon |
| Thesis advisor | Wetzstein, Gordon |
| Thesis advisor | Girod, Bernd |
| Thesis advisor | Horowitz, Mark (Mark Alan) |
| Degree committee member | Girod, Bernd |
| Degree committee member | Horowitz, Mark (Mark Alan) |
| Associated with | Stanford University, Department of Electrical Engineering |
Subjects
| Genre | Theses |
|---|---|
| Genre | Text |
Bibliographic information
| Statement of responsibility | David Brian Lindell. |
|---|---|
| Note | Submitted to the Department of Electrical Engineering. |
| Thesis | Thesis Ph.D. Stanford University 20210. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2020 by David Brian Lindell
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...