DIGITAL GEOLOGICAL MAPPING USING VALIDATED SUPERVISED CLASSIFICATION OF ASTER IMAGES IN THE SOROWAKO ULTRAMAFIC COMPLEX, SOUTH SULAWESI

• Main Author: Adika Nugraha, Yudhistira
• Format: Theses
• Language: Indonesia
• Subjects: Geologi, hidrologi & meteorologi
• Online Access: https://digilib.itb.ac.id/gdl/view/80850
• Institution: Institut Teknologi Bandung

The Sorowako Ultramafic Complex is one of the areas with a wide distribution of ultramafic rocks on Sulawesi Island with various geological potential contained therein. So far, detailed field mapping still has big challenges in the form of time and costs, especially for new areas with wide coverage. One method to determine the widespread distribution of rocks is to use Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) satellite image analysis because this image can delineate rock types from certain minerals spectrally. This research aims to identify the distribution of ultramafic rocks and other rocks visually using ASTER imagery validated by field observations. The method used is digital mapping using a spectral classification algorithm to obtain the best lithological classification for rocks exposed at the surface based on ASTER images validated with field data. In this study, three spectral classification methods were used, namelu Spectral Angle Mapper, Spectral Information Divergence, and Support Vector Machine, with the results of the Spectral Information Divergence classification method producing the best rock classification with an accuracy of 82.86% using a combination of 9 ASTER bands and 10 ASTER-derived images. Integration of ASTER imagery with field observations can effectively differentiate and classify rocks exposed in the Sorowako Ultramafic Complex by dividing them into nine rock units. The use of a supervised classification application using ASTER imagery can help identify the distribution of ultramafic rocks in the Sorowako area. The right combination of ASTER bands can provide effective classification results for field conditions with minimal vegetation even though the level of soil weathering is quite high.