The State of the Art of Modeling Millimeter-Wave
Remote Sensing of the Environment
KEVIN O'NEILL
sconced in sophisticated research structures. They
INTRODUCTION
may require advanced, specialized knowledge,
Recent improvements in technology and the
extremely complex, difficult-to-obtain input data,
prospect of superior performance have greatly
and intolerably large computational resources. At
increased interest in millimeter-wave (MMW) re-
least in extreme instances of this sort, such mod-
mote sensing. To assess the state of the art in the
els cannot be pursued here for the purposes of
accompanying modeling capabilities, a study was
ultimate scene generation. At the same time, the
undertaken with particular emphasis on im-
tension between generality and validity on the
plemented models, as opposed to disembodied
one hand and simplicity and accessibility on the
theory. This report summarizes the results of that
other may force a degree of compromise in priori-
study. Its aim is to survey all available MMW
tizing models. The requirements may have to be
modeling of scattering, emission, and propaga-
relaxed in some cases; in others, utility may be
tionabsorption in the natural environment, par-
achieved only by following an ongoing develop-
ticularly in the spectral vicinity of 35 and 94 GHz.
ment or working with model originators over an
The models are evaluated equally in terms of
extended period. While the field is developing,
validity and utility. In the context of this survey,
some applications may be best accomplished by
model means a theoretical formulation that is de-
combining the capabilities of theoretical models
rived from first principles, such as Maxwell's
with organized, facilitated databases. In the latter
equations or conservation of energy, and that is
connection, documentation and programs exist
incorporated in a tractable computational vehicle.
for extracting, manipulating, visualizing, and add-
Thus, some work that constitutes a model in the
ing to the large amount of measured data that
most general sense is recognized here for atten-
has evolved over recent decades--e.g., FINRACS
tion in the future, but it is not currently singled
(Borel et al. 1986) and material from Ulaby and
out as available through accessible software. The
his co-workers at the University of Michigan
ultimate purpose for identifying, collecting, and
(Ulaby and Dobson 1989a, 1989b; Ulaby and Had-
integrating an ensemble of such models is to gen-
dock 1990)
erate energy field distributions, i.e., scenes, that
The specific information incorporated in this
sensors might confront under the widest possible
survey is derived from a variety of sources: the
variety of conditions. To achieve this generality,
author's personal knowledge and information di-
models were sought based as much as possible
rect from the open literature, discussion and com-
on first principles, with minimal reliance on em-
munication with various recognized modelers, the
pirical or arbitrarily tunable parameterization.
response of modelers to a questionnaire, and a
Truly useful models should be sufficiently sim-
workshop (held on 1315 May 1992) for a general
ple and accessible that users other than the origi-
review and discussion of MMW modeling capa-
nators can run them. This requirement tests the
bilities. (A list of workshop participants appears
constraints of generality and first-principles va-
in Appendix A.) Beyond the specific topics for
lidity. Due to the nature of the field, we may find
the workshop sessions, far-ranging discussions
in many instances that the most rigorous models,
took place to establish the boundaries and con-
or perhaps the only rigorous models, are en-
tent of the current state of the art. In soliciting
Previous Page
