The Performance of Hyperspectral Measurements Integrated into Multispectral Data on the Detection of the Alpine Chromite Deposits in the Ophiolite Complexes

Abstract

Intense tectonic activities and alterations in ophiolites of the Alpine-Himalayan Orogeny Belt complicate the delineation and mapping of chromite-bearing rocks by traditional geology surveys. This study tests the performance of ASTER-based remote sensing in a region covered by ophiolite rocks with complex geology, alteration, and intense tectonism in Eastern-Central Anatolia. Today, there are many chromite deposits and detected mineralizations in the study area. The fact that there was almost no chromite mining activity in the study area at the acquisition date of the ASTER data, highlights the usability of the results of the remote sensing methods and techniques applied in the study for mineral exploration activities. According to field surveys, petrographic examinations, and ore microscopy, the ophiolite rocks consist of serpentinized peridotite, harzburgite, dunite, diabase, and gabbro. Chromite mineralizations occur in dunites such as disseminated, banded, and lens-shaped. Hyperspectral signatures obtained from the representative rock and mineralization samples using ASD FieldSpec 4 Hi-Res have been integrated into the advanced spaceborne thermal emission and reflection radiometer (ASTER) multispectral satellite data. This study shows the performance of image enhancement techniques such as band ratio, minimum noise fraction (MNF), and spectral classification such as matched filtering (MF). The boundaries of ophiolite rocks with band ratio techniques are distinguished at the regional scale of the boundaries of sedimentary and igneous rocks, but sub-ophiolite rock units are not distinctly appearing within the ophiolite rocks. The MNF results can distinguish chromite-bearing rocks from the other rocks, although the boundaries of the rocks are not very distinct. The MF results map the sub-ophiolite units with high accuracy, except for the diabase dykes. The MF results have an overall accuracy of over 84% and a kappa coefficient of 0.74. Most of the chromite-bearing sites identified in the MF are currently being operated as chromite deposits. Consequently, this study shows that the MF spectral classification can be high performance in the delineation of chromite-bearing rocks in the ophiolite complexes.