Exploring Gördes Zeolite Sites by Feature Oriented Principle Component Analysis of LANDSAT Images

Abstract

Recent studies showed that remote sensing (RS) is an effective, efficient and reliable technique used in almost all the areas of earth sciences. Remote sensing as being a technique started with aerial photographs and then developed employing the multi-spectral satellite images. Nowadays, it benefits from hyper-spectral, RADAR and LIDAR data as well. This potential has widen its applicability in the various areas and professional disciplines much more efficiently as never been before. One of the areas that remote sensing has been applied well and has become one of the indispensable tools for the earth science's scientists are geologic and mineral exploration studies and especially prospection stages of these studies. In this research, it was tried to determine and to map zeolite sites in Gördes region (Turkey) which were formed as alteration products having high level of water content and developed in volcanic rock beds by the help of remote sensing and GIS. The study area is about 400 km2 and located at the North-East of Manisa Province. The results confirmed that the zeolite areas obtained by classical exploration techniques can be determined using remote sensing techniques such as Feature Oriented Principal Component Analysis (PCA). Other zeolite areas in the same scene were also determined or at least predicted by this computer learning process through the same remote sensing image analyses.