|
Distinguishing Lithological Units Based on Satellite Imagery Using Machine Learning and Deep Learning Methods in the Hendudar Region, Markazi Province
|
Mehrtash Manafifard *   , Mostafa Yousefirad |
| Arak University of Technology |
|
|
Abstract: (14 Views) |
Remote sensing imagery is widely used in various geological applications, including the identification and discrimination of different rock types. In this study, artificial intelligence methods, including Random Forest, Deep Learning, XGBoost, and LightGBM, were employed to distinguish four geological units, Phyllite, Ancient Alluvial Terraces, Granite to Granodiorite, and Migmatite, in the Hendudar area, Markazi Province, Iran.
The input data for the AI models consisted of features extracted from satellite imagery such as Landsat, ASTER, and Sentinel-2, along with topographic data derived from a Digital Elevation Model (DEM), including elevation, slope, Topographic Wetness Index (TWI), and aspect. In addition, spatial data and field samples of the geological units were incorporated into the analysis.
To prepare the input datasets, necessary preprocessing steps, including radiometric and atmospheric corrections, as well as initial image processing and classification, were performed. Subsequently, a set of features was generated based on spectral bands of satellite imagery, as well as transformed datasets derived from Principal Component Analysis (PCA) and Minimum Noise Fraction (MNF) techniques.
The evaluation results indicate that the Random Forest algorithm achieved the best performance. Using spectral information from ASTER imagery combined with topographic and spatial data, the model attained Kappa, Overall Accuracy, and F-score values of 0.74, 0.82, and 0.82, respectively.
|
|
| Keywords: Geology, Lithological mapping, Satellite imagery, Machine learning, Deep learning |
|
|
|
Type of Study: Applicable |
Subject:
RS
Received: 2025/10/31 | Accepted: 2026/06/10 | ePublished ahead of print: 2026/06/17
|
|
|
|
|
|
|
| Send email to the article author |
|
|
