Estimating water and ice content on planetary soils using neutron measurements: A neural network approach

A. Luciani, P. Panfili, Roberto Furfaro, Barry D Ganapol, D. Mostacci

Research output: Contribution to journalArticle

Abstract

A model-based neural network methodology to estimate water and ice content in planetary soils using neutron fluxes detected by in situ and/or airborne deployment of neutron detectors is proposed and shown to be effective. Focusing of epithermal and thermal energy regimes, the neutron fluxes are computed [Panfili, P.; Luciani, A.; Furfaro, R.; Ganapol, B.D.; Mostacci, D. Radiat. Eff. Defects Solids 2009, 164 (5-6), 340-344.] as functions of the medium physical properties and used to train neural networks in the inverse mode. For homogeneous soil, the model-based neural network shows satisfactory performances in retrieving the percentage of water. For soil modelled as layered, neural networks designed to retrieve both the depth and thickness of an ice layer beneath the soil surface provide good results only in a limited range of configurations. However, it has been found that training the two networks to independently retrieve the two parameter results more accurately. It has also been found that multiple measurements help improve the accuracy of the inversion for this configuration.

Original languageEnglish (US)
Pages (from-to)345-349
Number of pages5
JournalRadiation Effects and Defects in Solids
Volume164
Issue number5-6
DOIs
StatePublished - May 2009

Fingerprint

Ice
moisture content
soils
Neutrons
ice
estimating
Neural networks
Soils
neutrons
Neutron flux
Water
flux (rate)
Neutron detectors
Thermal energy
neutron counters
configurations
thermal energy
Physical properties
education
Defects

Keywords

  • Inverse transport problem
  • Neural networks
  • Neutron flux
  • Neutron transport

ASJC Scopus subject areas

  • Condensed Matter Physics
  • Nuclear and High Energy Physics
  • Radiation
  • Materials Science(all)

Cite this

Estimating water and ice content on planetary soils using neutron measurements : A neural network approach. / Luciani, A.; Panfili, P.; Furfaro, Roberto; Ganapol, Barry D; Mostacci, D.

In: Radiation Effects and Defects in Solids, Vol. 164, No. 5-6, 05.2009, p. 345-349.

Research output: Contribution to journalArticle

@article{6a8200ba3d3a44268ca003d57aaebd2f,
title = "Estimating water and ice content on planetary soils using neutron measurements: A neural network approach",
abstract = "A model-based neural network methodology to estimate water and ice content in planetary soils using neutron fluxes detected by in situ and/or airborne deployment of neutron detectors is proposed and shown to be effective. Focusing of epithermal and thermal energy regimes, the neutron fluxes are computed [Panfili, P.; Luciani, A.; Furfaro, R.; Ganapol, B.D.; Mostacci, D. Radiat. Eff. Defects Solids 2009, 164 (5-6), 340-344.] as functions of the medium physical properties and used to train neural networks in the inverse mode. For homogeneous soil, the model-based neural network shows satisfactory performances in retrieving the percentage of water. For soil modelled as layered, neural networks designed to retrieve both the depth and thickness of an ice layer beneath the soil surface provide good results only in a limited range of configurations. However, it has been found that training the two networks to independently retrieve the two parameter results more accurately. It has also been found that multiple measurements help improve the accuracy of the inversion for this configuration.",
keywords = "Inverse transport problem, Neural networks, Neutron flux, Neutron transport",
author = "A. Luciani and P. Panfili and Roberto Furfaro and Ganapol, {Barry D} and D. Mostacci",
year = "2009",
month = "5",
doi = "10.1080/10420150902811656",
language = "English (US)",
volume = "164",
pages = "345--349",
journal = "Radiation Effects and Defects in Solids",
issn = "1042-0150",
publisher = "Taylor and Francis Ltd.",
number = "5-6",

}

TY - JOUR

T1 - Estimating water and ice content on planetary soils using neutron measurements

T2 - A neural network approach

AU - Luciani, A.

AU - Panfili, P.

AU - Furfaro, Roberto

AU - Ganapol, Barry D

AU - Mostacci, D.

PY - 2009/5

Y1 - 2009/5

N2 - A model-based neural network methodology to estimate water and ice content in planetary soils using neutron fluxes detected by in situ and/or airborne deployment of neutron detectors is proposed and shown to be effective. Focusing of epithermal and thermal energy regimes, the neutron fluxes are computed [Panfili, P.; Luciani, A.; Furfaro, R.; Ganapol, B.D.; Mostacci, D. Radiat. Eff. Defects Solids 2009, 164 (5-6), 340-344.] as functions of the medium physical properties and used to train neural networks in the inverse mode. For homogeneous soil, the model-based neural network shows satisfactory performances in retrieving the percentage of water. For soil modelled as layered, neural networks designed to retrieve both the depth and thickness of an ice layer beneath the soil surface provide good results only in a limited range of configurations. However, it has been found that training the two networks to independently retrieve the two parameter results more accurately. It has also been found that multiple measurements help improve the accuracy of the inversion for this configuration.

AB - A model-based neural network methodology to estimate water and ice content in planetary soils using neutron fluxes detected by in situ and/or airborne deployment of neutron detectors is proposed and shown to be effective. Focusing of epithermal and thermal energy regimes, the neutron fluxes are computed [Panfili, P.; Luciani, A.; Furfaro, R.; Ganapol, B.D.; Mostacci, D. Radiat. Eff. Defects Solids 2009, 164 (5-6), 340-344.] as functions of the medium physical properties and used to train neural networks in the inverse mode. For homogeneous soil, the model-based neural network shows satisfactory performances in retrieving the percentage of water. For soil modelled as layered, neural networks designed to retrieve both the depth and thickness of an ice layer beneath the soil surface provide good results only in a limited range of configurations. However, it has been found that training the two networks to independently retrieve the two parameter results more accurately. It has also been found that multiple measurements help improve the accuracy of the inversion for this configuration.

KW - Inverse transport problem

KW - Neural networks

KW - Neutron flux

KW - Neutron transport

UR - http://www.scopus.com/inward/record.url?scp=67749117924&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67749117924&partnerID=8YFLogxK

U2 - 10.1080/10420150902811656

DO - 10.1080/10420150902811656

M3 - Article

AN - SCOPUS:67749117924

VL - 164

SP - 345

EP - 349

JO - Radiation Effects and Defects in Solids

JF - Radiation Effects and Defects in Solids

SN - 1042-0150

IS - 5-6

ER -