Abstract
The work describes various methods employed towards solving the problem of indirect imaging. Computational techniques are employed to indirectly decipher information about an object hidden from view of a camera. Notion of virtualizing the source of illumination and detectors on real world rough surfaces was exploited to construct a non line of sight computational imager. Diversity was explored from the stand point of both illumination of the object and imaging of light reflected from the object. To understand the impact of scattering by real world rough surfaces, an instrument was developed that allows characterization of isoplanatic angle for different surface types. Various aspects (impact of absorption, multiple scatter, roughness scales, etc) of scattering from a surface was explored and identified. A computational scheme was identified to isolate the contribution of singly scattered light from multiply scattered light by employing techniques borrowed from linear algebra. An experimental testbed that uses continuous wave(CW) sources to create spatially resolved intensity images of objects completely hidden from line of sight was developed. Ideas were borrowed from imaging correlography, a line of sight imaging technique widely used in satellite imaging. Experiments were carried out to explore limits of this non-line-of-sight computational imager. A testbed was constructed that could convert real world rough surfaces into virtualized pattern projectors in order to illuminate the hidden object with known light patterns. A mathematical model was partially developed to understand the capability of identifying depth and 3D information of the hidden object using the virtualized pattern projector. The idea of moving the center of perspective of the imaging system to the plane of the wall was proposed by moving the entrance pupil of the imager to the plane of the wall. This could potentially endow the system with stereo imaging capability. A Zemax simulation was carried out to identify the feasibility of this idea.
Degree Date
Fall 2018
Document Type
Thesis
Degree Name
M.S.E.E.
Department
Electrical and Computer Engineering
Advisor
Marc Christensen
Second Advisor
Prasanna Rangarajan
Third Advisor
Duncan MacFarlane
Subject Area
Physical Sciences, Physics
Format
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License
Recommended Citation
Viswanath, Aparna, "Indirect Imaging Using Computational Imaging Techniques" (2018). Electrical Engineering Theses and Dissertations. 17.
https://scholar.smu.edu/engineering_electrical_etds/17
