University of Twente Student Theses


Methodology for quantifying the usefulness of remotely sensed data in studying geology, a case study in the Cabo de Gata, southeast Spain

Hung, Ha Quang (2010) Methodology for quantifying the usefulness of remotely sensed data in studying geology, a case study in the Cabo de Gata, southeast Spain.

[img] PDF
Abstract:The aim for this research is to determine what type of geological features can be extracted from remotely sensed data and to develop a methodology for quantifying the usefulness of remotely sensed data in studying geology. Four different types of remotely sensed data (ASTER, ASTER-derived DEM, HyMap and ALOS PALSAR) were used in studying geology in the Cabo de Gata, southeast Spain. Field data were used in combining with remotely sensed data to define rules for image classification. Typical characteristics for the discrimination of lithological formations based on field properties and images properties were organized in a columnal display by mean of spectral / remotely sensed stratigraphic column. Defined characteristics for recognizing lithological formations as shown in the spectral / remotely sensed stratigraphic column were used to interpret available remotely sensed data. The usefulness of remotely sensed data in studying geology was quantified by estimating the goodness-of-fit between remotely sensed classification and published geological maps and measuring the differences between actual agreement and chance agreement with Kappa statistics. Some notices are summarized as below:  Reflectance spectra of 180 field samples were analysed. Surface minerals obtained from reflectance spectra were used to define end-members for the image classification.  66 gamma-ray measurements were made covering five lithologies in Carboneras fault zone. Radioactive signatures of lithologies were grouped into categories using normalized Ternary Radioement Mapping and statistical analysis with boxplot.  Band-ratioing and false colour composite were chosen to process ASTER and HyMap images for visual interpretation. False colour composites of R: band-ratio 4/1; G: 3/1 and B: 12/14 and R: 1648/558; G: 816/558 and B: 1597/2205 were created for ASTER and HyMap images repetitively. Lithologies were recognized and grouped into categories based on their false colour composite expressions.  Spectral Angle Mapping (SAM) and Spectral Feature Fitting (SFF) were used as mapping methods to process ASTER and HyMap images. Images classified with SAM were chosen as optimal end-products obtained from automatic classification. Supervised classification (Maximum Likelihood Classification) was used to process ALOS PALSAR image with prior knowledge about geology of the study area.  According to KHAT statistics, the differences between observed classification and the classification resulted from chance of various remotely sensed data were ranked as follow: HyMap (64.92%) - ALOS PALSAR (40.96%) - ASTER combines with ASTER-derived DEM (38.15%) - and ASTER (21.93%). Keywords: Remotely sensed data; Cabo de Gata; geologic remote sensing; usefulness; spectral / remotely sensed stratigraphic column; Kappa statistics; ASTER; ASTER-derived DEM; HyMap; ALOS PALSAR; band-ratio; false colour composite; rule-based classification.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page