Hutagalung, Daniel Sahat Rezeki (2024) Pemetaan Litologi Dengan Data Gayaberat Satelit Dan Citra Satelit Menggunakan Sampel Spasial Dan Algoritma Convolutional Neural Network. Other thesis, Institut Teknologi Sepuluh Nopember.
Text
5017201021-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (5MB) | Request a copy |
Abstract
Berkembangnya teknologi citra satelit dapat digunakan untuk menghasilkan peta litologi atau mendetailkan peta yang sudah ada. Salah satu cara memanfaatkan data citra satelit dan data geofisika satelit dalam pemetaan litologi adalah dengan menggunakan algoritma Convolutional Neural Networks (CNN). Meskipun algortitma CNN efektif dalam pengolahan data citra, optimalisasi untuk pemetaan litologi memerlukan teknik pengambilan sampel spasial yang lebih baik untuk meningkatkan akurasi dalam menginterpretasikan data geospasial yang kompleks. Algoritma CNN digunakan pada data spasial berupa Landsat 8, DEM, dan Gayaberat satelit sebagai data fitur dan sampel data geologi inderaan jauh sebagai data label untuk memprediksi litologi permukaan daerah penelitian. Validasi dari hasil pemetaan menggunakan confusion matrix dan peta geologi inderaan jauh daerah penelitian. Dengan teknik sampel spasial random, systematic centric, dan cluster, dengan variasi jumlah titik sampel 40, 84, dan 160 titik yang mencakup 10% dari peta geologi inderaan jauh. Hasil model menunjukkan akurasi sampel random sebesar 0.91 – 0.937, systematic centric sebesar 0.726 – 0.847, dan cluster sebesar 0.861 – 0.928. Model yang memiliki akurasi tertinggi diperoleh dari sampel random dengan 84 titik, yaitu sebesar 0.937.
==============================================================================================================================
The development of satellite imagery technology can be used to produce lithological maps or to detail existing maps. One way to utilize satellite imagery data and satellite geophysical data in lithological mapping is by using Convolutional Neural Networks (CNN) algorithms. Although CNN algorithms are effective in processing image data, optimizing for lithological mapping requires better spatial sampling techniques to improve accuracy in interpreting complex geospatial data. The CNN algorithm is used on spatial data, including Landsat 8, DEM, and satellite gravity as feature data, and remote sensing geological data as label data to predict the surface lithology of the study area. The mapping results are validated using a confusion matrix and the remote sensing geological map of the study area. With spatial sampling techniques including random, systematic centric, and cluster sampling, and varying the number of sample points to 40, 84, and 160 points, which cover 10% of the geological map, the results show that the accuracy of the random sample is 0.91 – 0.937, the systematic centric sample is 0.726 – 0.847, and the cluster sample is 0.861 – 0.928. The model with the highest accuracy was obtained from a random sample with 84 points, achieving an accuracy of 0.937.
Item Type: | Thesis (Other) |
---|---|
Uncontrolled Keywords: | Complete Bouger Anomaly, Convolutional Neural Network, Landsat-8, Lithology Mapping, Spatial Sampling, Pemetaan Litologi, Sampel Spasial |
Subjects: | G Geography. Anthropology. Recreation > GA Mathematical geography. Cartography > GA102.4.R44 Cartography--Remote sensing |
Divisions: | Faculty of Civil, Planning, and Geo Engineering (CIVPLAN) > Geophysics Engineering > 33201-(S1) Undergraduate Thesis |
Depositing User: | Daniel Sahat Rezeki Hutagalung |
Date Deposited: | 21 Aug 2024 02:55 |
Last Modified: | 21 Aug 2024 02:55 |
URI: | http://repository.its.ac.id/id/eprint/114198 |
Actions (login required)
View Item |