A new image fusing approach for mineral mapping

Except for the short-wave infrared data, thermal infrared (TIR) wavelengths have an important in geological remote sensing. It is possible to distinguish quartz and evaporate minerals using TIR data. The most important data source for thermal infrared imaging is the ASTER satellite sensor. However, due to its nature, the spatial resolution of TIR data is low. Therefore, its use in detailed mineral mapping is limited. The aim of this study is to present a new approach to fuse ASTER thermal satellite images (TIR), which are frequently used in geological surveys, with visible near infrared (VNIR) data. For this purpose, the proposed approach was tested in the region rich in gypsum minerals in Bala district of Ankara. In the study carried out, gypsum minerals of the region were mapped with the proposed method (FVT). In addition, vegetation, water and other lithological units in the region were examined with the proposed approach. The results obtained were compared in terms of statistical and spatial performances. In the evaluation, the differences in the visual, spectral signature, spectral angle and sulphate indices of the original thermal data and the proposed approach were examined. The optimum threshold value of TIR and FVT data was taken and the proposed approach showed minimal difference, especially in regions with gypsum. The field studies of the mapped gypsum units were checked with the satellite images and UAV images taken from the region were compared, it was seen that the proposed approach had higher accuracy. In addition, chemical analysis of gypsum samples taken from the region was made and it was determined that the areas classified as gypsum in the study contained 99.3% gypsum.