Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/123254
Title: Deteksi Perubahan Penggunaan dan Penutupan Lahan Menggunakan Machine Learning pada Google Earth Engine
Other Titles: Detection of Land Use and Land Cover Changes in Sukajaya District Using Machine Learning on Google Earth Engine
Authors: Tjahjono, Boedi
Munibah, Khursatul
Rivai, Fathan Aldi
Issue Date: 8-Aug-2023
Publisher: IPB University
Abstract: Perubahan penggunaan/penutupan lahan merupakan fenomena yang terjadi hampir di seluruh bagian dunia. Beberapa wilayah telah terjadi perubahan penggunaan/penutupan lahan yang intens, terutama di Kecamatan Sukajaya, Kabupaten Bogor. Tujuan penelitian ini adalah menelaah algoritma indeks, membandingkan machine learning dalam mengklasifikasi penggunaan/penutupan lahan di Kecamatan Sukajaya, Kabupaten Bogor, serta analisis perubahannya dari tahun 2013-2022. Kelas penggunaan/penutupan lahan dibagi menjadi tujuh kelas. Algoritma indeks yang digunakan sebanyak 12 indeks. Machine learning yang digunakan adalah CART, Random Forest dan SVM. Enhanced Water Index adalah indeks yang paling dapat membedakan pengunaan/penutupan lahan di Kecamatan Sukajaya. SVM performa terbaik dengan nilai OA 90,46 dan Kappa 88,80 di tahun 2013 dan nilai OA 91,08 dan Kappa 89,52 di tahun 2022. Sifat irreversible perluasan lahan terbangun di Kecamatan Sukajaya tidak mutlak karena telah terjadi longsor yang menyebabkan perubahan klasifikasi lahan terbangun menjadi lahan terbuka.
Land use/cover change is a phenomenon that occurs in almost every part of the world. Some regions have experienced intense changes in land use/cover, especially in Sukajaya District, Bogor Regency. The purpose of this study is to examine index algorithms, compare machine learning methods in classifying land use/cover in Sukajaya District, Bogor Regency, and analyze the land use/land cover changes from 2013 to 2022. Land use/cover classes are divided into seven categories. A total of 12 index algorithms are used. The machine learning methods used are CART, Random Forest, and SVM. The Enhanced Water Index is the best index in differentiating land use/cover in Sukajaya District. SVM show the best performance with an Overall Accuracy (OA) of 90.46% and Kappa value of 88.80% in 2013, and an OA of 91.08% and Kappa value of 89.52% in 2022. The irreversible characteristic of built-up area expansion in Sukajaya District is not absolute due to a landslide that led to a change in the classification of built-up area to bare land.
URI: http://repository.ipb.ac.id/handle/123456789/123254
Appears in Collections:UT - Soil Science and Land Resources

Files in This Item:
File Description SizeFormat 
cover.pdf
  Restricted Access
Cover792.03 kBAdobe PDFView/Open
full teks.pdf
  Restricted Access
full text2.01 MBAdobe PDFView/Open
lampiran.pdf
  Restricted Access
lampiran175.87 kBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.