عنوان مقاله [English]
Exploration of the new mineral deposits around the existing mines is an important objective in mining industry. Using multispectral remote sensing images, due to their diversity and vast availability is a useful tool to meet this purpose. In this research, the Zonouz region was investigated for discovering new high potential kaolinite mineralization areas using Landsat8 and ASTER data. Zonouz kaolin mine which is located in Marand county, East-Azarbaijan, is the biggest kaolin deposit in the Middle-East. In the current research, the capability of Landsat8, as a new generation multispectral data, and ASTER data were examined in mineral detection. At first, the preprocessing of data, i.e. atmospheric and topographic corrections and elimination of the vegetation cover were carried out. Then, the spectral profiles of the endmembers of the study area datasets were extracted. Identification of the extracted endmembers was done by comparison of the unknown spectra with reference spectra of the USGS spectral library, and 3 minerals including kaolinite, quartz, and Fe-bearing minerals were identified. Finally, the distribution maps of the identified minerals were extracted by using of the artificial neural networks, as a non-linear supervised method. To the best of our knowledge, the applied neural networks structure has not been implemented on LANDSAT 8 and ASTER data earlier. This research is also the first implementation of LANDSAT 8 and ASTER data in Zonouz kaolin region for extension mapping of the kaolin. Two different approaches including virtual verification and field sampling were applied for validation of the results. According to the Findings of this research, both of ASTER and Landsat8 datasets proved successfulness for identifying the kaolinite; but the Landsat 8 data exhibited better performance in detecting and mapping of the quartz and hematite. Finally, 6 promising areas were determined as high potential zones of kaolinite mineralization for future studies.