Hiperspektral anomali ve hedef tespiti algoritmalarının çevresel uygulamaları.

The objective of this thesis study is to understand the capacity of hyperspectral anomaly detection and target detection algorithms for contamination mapping, with a focus on developing a new methodology for environmental problems related to coal mining operations. Considering the data availability, specific chemical structure and reflectance properties as well as being one of the common contaminants in mining operations, hydrocarbon induced problems are studied for water and soil medium separately. Having developed the algorithms and implemented new methodologies, the research is focused on understanding the potential of anomaly detection for identifying topographical changes in temporal manner and mapping secondary iron minerals related to acid mine drainage, which is a commonly encountered problem in coal mine sites. On the top of the developed expertise of the definitions and mathematical implications of anomaly and target detection algorithms, soils’ spectral features with regard to their chemical and physical relation is elaborated. The research contributes to literature by presenting the success of hyperspectral anomaly and target analysis algorithms to differentiate oil slicks in water medium. In addition to the capacity of anomaly and target analysis to determine hydrocarbon induced alterations, unmixing techniques are also proved to resolve the characteristic features in relation with surface manifestations of hydrocarbon seeps by means of multispectral data sets.This novel approach is adapted to determine the iron oxide mineral contents associated with acid mine drainage problems in mining areas with one of the latest imaging missions of European Space Agency (ESA), Sentinel-2. The theoretical background and experiments on synthetic data to map iron related features so far are combined to analyze the downloaded Sentinel-2 image with the help of the developed methodologies.Ph.D. - Doctoral Progra