Classification of geological mapping features using satellite remote sensing and in-situ spectro-radiometric measurements over Cyprus

Abstract

This paper aims at establishing the spectral reflectance signature for a number of geological mapping features and specific rocks over the area of Cyprus. This will enable the investigation for specific geological features through classification using satellite images. The purpose is to provide a useful tool for geologists in observation of surface strata. Methodology followed includes extraction of the spectral reflectance signature of the geological features by using satellite imagery, such as those of Landsat TM/ETM+, ASTER etc. In addition in-situ spectro-radiometric measurements were collected for the same feature locations. The selected sites included mines and quarries, with no vegetation cover and therefore no influence on results. Spectral reflectance for each feature refers to average value of retreated satellite image value and measurement result. An algorithm is finally established, aiming to be used for classification purposes of geological mapping and other applications. This innovated approach will, also, prove by validation the accuracy of each method for the spectral reflectance signature estimation. This additional benefit would conclude recommendation for future satellite sensors navigation and work processes. NIR band was found to be suitable for discriminating betonite, limestone and diabase geological features (as found at quarries and mines)