Evaluating the effects of spatial resolution on hyperspectral fire detection and temperature retrieval

Download item | Update item information
Publication Type thesis
School or College College of Social & Behavioral Science
Department Geography
Author Matheson, Daniel Scott
Title Evaluating the effects of spatial resolution on hyperspectral fire detection and temperature retrieval
Date 2011-08
Description Hyperspectral remote sensing of wildfires combines principles of emitted radiation with advanced spectrometry to model wildfire area and temperature, as well as background land cover classification, at the subpixel level. Yet airborne hyperspectral sensors face problems of inconsistent spatial resolutions and have limited spatial and temporal coverage. A proposed hyperspectral/thermal infrared satellite, the Hyperspectral InfraRed Imager (HyspIRI), will provide hyperspectral data over a spectral range of 350- 2500 nm at a spatial resolution of 60.0 m. Hyperspectral radiance data have previously been shown to allow fire detection and retrieval of fire temperature, although these abilities have not been demonstrated at spatial resolutions coarser than 16.1 m. For this study, four hyperspectral images containing active fires were acquired by the Airborne Visible-Infrared Imaging Spectrometer (AVIRIS), with spatial resolutions ranging from 3.8 to 16.1 m. By resampling these AVIRIS images to coarser spatial resolutions and by modeling fire area, fire temperatures and background land cover, the impacts of spatial resolution on fire detection and temperature retrieval were simulated. Multiple endmember spectral mixture analysis (MESMA) methods were used to model fire temperature and background land cover types. Modeling at coarser spatial resolutions produced larger areas of low fire temperatures with lower modeling error than modeling at finer spatial resolutions. Modeling results comparing 60.0 m data with and without a Gaussian point spread function validated pixel aggregation resampling as a suitable approximation of coarser spatial resolution imagery. Coarser spatial resolution hyperspectral data, such as that collected by the future HyspIRI sensor, are likely to model more fire area and lower temperatures when compared against simultaneously acquired higher spatial resolution data. Increasing the saturation thresholds of SWIR channels could greatly improve the fire detection and temperature modeling capabilities of a HyspIRI-like sensor.
Type Text
Publisher University of Utah
Subject AVIRIS; Hyperspectral; Remote sensing; Spatial scaling; Spectral mixture analysis; Wildfires
Dissertation Institution University of Utah
Dissertation Name Master of Science
Language eng
Rights Management Copyright © Daniel Scott Matheson 2011
Format Medium application/pdf
Format Extent 5,304,271 bytes
Identifier us-etd3,40270
Source Original housed in Marriott Library Special Collections, G59.5 2011 .M38
ARK ark:/87278/s6jt0552
Setname ir_etd
Date Created 2012-04-24
Date Modified 2017-11-02
ID 194640
Reference URL https://collections.lib.utah.edu/ark:/87278/s6jt0552
Back to Search Results