Optimizing Alteration Mineral Detection: A Fusion of Multispectral and Hyperspectral Remote Sensing Techniques in the Sar-e-Chah-e Shur, Iran

The Sar-e-Châh-e-Shur region in the Lut Block of eastern Iran has significant potential for diverse ore mineral deposits, particularly copper and iron mineralization. This study used multispectral remote sensing data, including ASTER, Sentinel 2, and Landsat 8 for detecting alteration zones associated with ore mineralization in the study area. Spectral ratios associated with porphyry copper-iron alterations and advanced argillic-iron alterations were utilized to establish training data, supplemented by hyperspectral data processing derived from Hyperion (spaceborne) and HyMap (airborne) sensors. Several classification algorithms were implemented, and evaluation metrics such as the Kappa coefficient and overall accuracy were calculated using a confusion matrix to assess classification performance. Field validation was conducted through thin mineral section analysis and geochemical stream sediment data, yielding a Normalized Score (NS) for conformity estimation. Notably, the Landsat 8, ASTER, and Sentinel 2 - classes category 2 -artificial neural networks-thresholds 2 (LAS-2-ANN2) class (representing advanced argillic alterations) was achieved a high score of 3.5, corresponding to an 87.5% match to the geological evidence. In addition, hyperspectral data processing using artificial neural network algorithms was yielded an overall score of 2.43, indicating a 61% match. In conclusion, this investigation highlighted the efficacy of fusing multispectral and hyperspectral data for accurate mineral alteration mapping in high potential regions, offering valuable insights for future exploration campaigns.