EVALUATING INTEGRATED OBJECT-BASED IMAGE ANALYSIS (OBIA) FOR SEMI-AUTOMATED LITHOLOGICAL FORMATIONS AND STRUCTURES DETECTION

Geologists have always depended on terrain visits and manual techniques to detect geological formations and tectonic structures to conduct geologic mapping studies; however, these classical methods remain time consuming, subjective, and unreliable. As such, semi-automated satellite-borne remote sensing data and image classification methods and approaches can reduce the subjectivity. Object-Based Image Analysis (OBIA) has become a mainstream technique for landuse and landcover (LULC) classification, however, they can also be applied for geological formation identification. In this work, a Sentinel-2A band combination was used as input for the Machine Learning powered Orfeo Toolbox (OTB) algorithms under QGIS to extract and digitize lithological formations. Obtained results provided a good discrimination of the lithological units for lithological mapping in comparison with the existing geological and field investigation. Furthermore, this geospatial technique remains a powerful tool that can help geologists carry out thorough geologic surveys in a modern way.