Abstract

The Digital Image and Remote Sensing Image Generation (DIRSIG) model has been developed and utilized to support research at the Rochester Institute of Technology (RIT) for over a decade. The model is an established, first-principles-based scene simulation tool that has been focused on passive multi- and hyper-spectral sensing from the visible to long wave infrared (0.4 to 14 pm). Leveraging photon mapping techniques utilized by the computer graphics community, a first-principles-based elastic Light Detection and Ranging (LIDAR) model was incorporated into the passive radiometry framework so that the model calculates arbitrary, time-gated photon counts at the sensor for atmospheric, topographic, and backscattered returns. The active LIDAR module handles a wide variety of complicated scene geometries, a diverse set of surface and participating media optical characteristics, multiple bounce and multiple scattering effects, and a flexible suite of sensor models. This robust modeling environment allows the researcher to evaluate sensor design trades for topographic systems and the impact that scattering constituents (e.g. water vapor, dust, sediment, soot, etc.) may have on a Differential Absorption LIDAR (DIAL) system's ability to detect and quantify constituents of interest within volumes including water and atmospheric plumes.

The interest in modeling DIAL sensor engagements involving participating media such as gaseous plumes presented significant challenges that were overcome using the photon mapping paradigm. Intuitively, researchers suspect that multiple scattering effects from additional constituents as simple as water vapor or soot could impact a DIAL sensor's ability to detect and quantify effluents of interest within a participating medium. Traditional techniques, however, are not conducive to modeling the multiple scattering and absorption within a non-homogenous finite volume, such as a plume. Specific numerical approaches are presented for predicting sensor-reaching photon counts, including the effects of multiple scattering and absorption within a realistic plume. These approaches are discussed and benchmarked against analytically predicted results using a non-stationary, diffusion approximation. The analytical development and consistency of this new variant of photon mapping is explored along with the underlying physics and radiative transfer theory for participating media. Additionally, a LIDAR equation that accounts for multiple scattering effects is presented in conjunction with a discussion of the importance of accounting for these effects.

Representative datasets generated via DIRSIG for both a topographical LIDAR and DIAL system are then shown. The results from some interesting phenomenological case studies including standard terrain topography, forest canopy penetration, plume interrogation with scattering and absorbing constituents, and camouflaged hard targets are also presented. Based upon a limited number of case studies, the effects of multiple scattering on DIAL sensor performance are also qualitatively discussed.

Library of Congress Subject Headings

Air--Pollution--Measurement--Optical methods; Remote sensing--Mathematical models; Electromagnetic waves--Scattering--Mathematical models; Light absorption--Mathematical models; Plumes (Fluid dynamics)

Publication Date

2005

Document Type

Dissertation

Student Type

Graduate

Degree Name

Imaging Science (Ph.D.)

Department, Program, or Center

Chester F. Carlson Center for Imaging Science (COS)

Advisor

John R. Schott

Advisor/Committee Member

Roger Easton

Advisor/Committee Member

John Kerekes

Comments

Physical copy available from RIT's Wallace Library at TD890 .B54 2005

Campus

RIT – Main Campus

Download

Share

COinS