Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Pays Bas → Lithological mapping of the Bas Al Hickman and Shuaiba carbonate facies using hyperspectral data in Oman

International Institute for Geo-Information Science and Earth Observation (ITC) 2003

Lithological mapping of the Bas Al Hickman and Shuaiba carbonate facies using hyperspectral data in Oman

Deen, Henry M.C

Titre : Lithological mapping of the Bas Al Hickman and Shuaiba carbonate facies using hyperspectral data in Oman

Auteur : Deen, Henry M.C

Etablissement de soutenance : University of Twente International Institute for Geo-Information Science and Earth Observation (ITC)

Grade : Master of Science in Geo-Information Science and Earth Observation 2003

Résumé
Remote sensing has undergone tremendous development in the past years and one such area which has been improved is the area of Hy- perspectral Imaging. Hyperspectral Imaging refers to the imaging of a scene over a large number of discrete, narrow contiguous spectral bands throughout the electromagnetic spectrum. The Barr Al Hick- man area is a prime example of a low relief modern carbonate ramp similar in many aspects to the geometries and facies distribution of the cretaceous carbonate reservoirs in Oman particularly the Shuaiba for- mation. This research investigates how hyperspectral mapping tech- niques such as Spectral Angle Mapper (SAM) and Linear Spectral Un- mixing performs in mapping the carbonate facies onshore and the un- derstanding of the spatial arrangement of the coral and substrate fa- cies in the offshore Barr Al Hickman area. The HyMAP data was acquired already pre-processed using Para- metric geocoding algorithm and a Radiative transfer Code (MODTRAN) to remove atmospheric effects from the data and to derive surface re- flectance values. Processing of the onshore component starts by select- ing pure end-members from the image using Minimum Noise Trans- formation, Pixel Purity Index and N-dimension visualizer. The im- age end-member spectra for halite, gypsum, borax, sepiolite, calcite, dolomite and kaolinite were identified by comparison with spectra from JPl library and were then used as input for classification using SAM and Linear Unmixing. In the processing of the offshore data compo- nent water column correction was done to compensate for variable wa- ter depth by calculating the ratio of attenuation coefficients for band pairs and subsequent generation of depth invariant bottom images. Series of colour composite images and transects were generated as simulated Landsat TM images to enhance visual interpretation and simulation of ground truth for classification. End-member spectra for substrates types (living corals, brown algae and sand) were used as inputs for SAM classification. Though no ground data was available for performing detailed accu- racy assessment a proposed validation method for the SAM algorithm shows that the results gives a better understanding of the mineral dis- tribution as compared to Linear spectral unmixing. It also proves to be useful in mapping the offshore coral reef substrates and facies. In a nutshell the hyperspctral data (HyMAP) demonstrated much higher seperability and spectral variation and the results suggests the use- fulness of airborne hyperspectral data to the study of offshore and on- shore carbonate facies distribution.

Version intégrale (ITC)

Page publiée le 31 janvier 2018