Show simple item record

dc.contributor.advisorChairperson, Graduate Committee: Kevin S. Repaskyen
dc.contributor.authorMcCann, Cooper Patricken
dc.date.accessioned2017-10-10T21:22:28Z
dc.date.available2017-10-10T21:22:28Z
dc.date.issued2017en
dc.identifier.urihttps://scholarworks.montana.edu/xmlui/handle/1/12799en
dc.description.abstractLow-cost flight-based hyperspectral imaging systems have the potential to provide valuable information for ecosystem and environmental studies as well as aide in land management and land health monitoring. This thesis describes (1) a bootstrap method of producing mesoscale, radiometrically-referenced hyperspectral data using the Landsat surface reflectance (LaSRC) data product as a reference target, (2) biophysically relevant basis functions to model the reflectance spectra, (3) an unsupervised classification technique based on natural histogram splitting of these biophysically relevant parameters, and (4) local and multi-temporal anomaly detection. The bootstrap method extends standard processing techniques to remove uneven illumination conditions between flight passes, allowing the creation of radiometrically self-consistent data. Through selective spectral and spatial resampling, LaSRC data is used as a radiometric reference target. Advantages of the bootstrap method include the need for minimal site access, no ancillary instrumentation, and automated data processing. Data from a flight on 06/02/2016 is compared with concurrently collected ground based reflectance spectra as a means of validation achieving an average error of 2.74%. Fitting reflectance spectra using basis functions, based on biophysically relevant spectral features, allows both noise and data reductions while shifting information from spectral bands to biophysical features. Histogram splitting is used to determine a clustering based on natural splittings of these fit parameters. The Indian Pines reference data enabled comparisons of the efficacy of this technique to established techniques. The splitting technique is shown to be an improvement over the ISODATA clustering technique with an overall accuracy of 34.3/19.0% before merging and 40.9/39.2% after merging. This improvement is also seen as an improvement of kappa before/after merging of 24.8/30.5 for the histogram splitting technique compared to 15.8/28.5 for ISODATA. Three hyperspectral flights over the Kevin Dome area, covering 1843 ha, acquired 06/21/2014, 06/24/2015 and 06/26/2016 are examined with different methods of anomaly detection. Detection of anomalies within a single data set is examined to determine, on a local scale, areas that are significantly different from the surrounding area. Additionally, the detection and identification of persistent anomalies and non-persistent anomalies was investigated across multiple data sets.en
dc.language.isoenen
dc.publisherMontana State University - Bozeman, College of Letters & Scienceen
dc.subject.lcshCarbon dioxideen
dc.subject.lcshSpectrum analysisen
dc.subject.lcshAerial photographyen
dc.subject.lcshTestingen
dc.titleMesoscale, radiometrically referenced, multi-temporal hyperspectral data for CO 2 leak detection by locating spatial variation of biophysically relevant parametersen
dc.typeDissertationen
dc.rights.holderCopyright 2017 by Cooper Patrick McCannen
thesis.degree.committeemembersMembers, Graduate Committee: Scott Powell; Wm. Randall Babbitt; Rufus L. Cone; Rand Swanson.en
thesis.degree.departmentPhysics.en
thesis.degree.genreDissertationen
thesis.degree.namePhDen
thesis.format.extentfirstpage1en
thesis.format.extentlastpage194en
mus.data.thumbpage101en


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record


MSU uses DSpace software, copyright © 2002-2017  Duraspace. For library collections that are not accessible, we are committed to providing reasonable accommodations and timely access to users with disabilities. For assistance, please submit an accessibility request for library material.