Ilyas, Ilyas (2020) Analisis Klasifikasi Citra Satelit WorldView-2 Menggunakan Model Deep Learning (Studi Kasus : Peta Tutupan Lahan Perkotaan Terangun, Kabupaten Gayo Lues, Provinsi Aceh). Masters thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
03311750012004-Master_Thesis.pdf Download (11MB) | Preview |
Abstract
Dalam menyusun sebuah dokumen Rencana Detil Tata Ruang (RDTR) maka diperlukan berbagai informasi peta dasar dari suatu wilayah yang akan direncanakan. Salah satu informasi tersebut adalah tutupan lahan. Informasi tutupan lahan merupakan informasi tahap awal yang harus tersedia agar para penyusun dokumen tata ruang memahami kondisi keruangan pada suatu area perencanaan. Informasi tutupan lahan diperoleh dari hasil klasifikasi citra satelit WorldView-2 dengan resolusi spasial 0,5 meter. Penerapan dari klasifikasi model deep learning yaitu peta tutupan lahan perkotaan Terangun, Kabupaten Gayo Lues, Provinsi Aceh dengan skala 1:5.000 yang terdiri dari bangunan, jalan, lahan terbuka, sungai, sawah, ladang, kolam, dan kebun. Uji akurasi klasifikasi model deep learning ini menunjukkan akurasi 65% overall accuracy dan 48,09% kappa accuracy, sedangkan metode maximum likelihood classification menunjukkan 49% overall accuracy dan dan kappa accuracy sebesar 24,56%. Sehingga dapat disimpulkan klasifikasi model deep learning mampu melakukan klasifikasi lebih baik dibandingkan dengan metode maximum likelihood classification.
====================================================================================================================================
In compiling a detailed Spatial Planning (RDTR) document, various basic map information from an area is needed. One such information is land cover. Land cover information is the initial information that must be available so that the drafters of spatial documents understand the spatial conditions in a planning area. Land cover information was obtained from the classification of WorldView-2 satellite imagery with a spatial resolution of 0.5 meters. The application of the deep learning model classification is a map of urban land cover Terangun, Gayo Lues Regency, Aceh Province with a scale of 1: 5,000 consisting of buildings, roads, open land, rivers, rice fields, fields, ponds and gardens. The classification accuracy test of this deep learning model shows the accuracy of 65% overall accuracy and 48.09% kappa accuracy, while the maximum likelihood classification method shows 49% overall accuracy and and kappa accuracy of 24.56%. So it can be concluded that the classification of deep learning model is able to do a better classification than the maximum likelihood classification method.
| Item Type: | Thesis (Masters) |
|---|---|
| Additional Information: | RTG 621.367 8 Ily a-1 |
| Uncontrolled Keywords: | Artificial Intelligence, Deep Learning, Detail Spatial Planning, Landcover, WorldView-2. |
| Subjects: | T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques. Image analysis--Data processing. |
| Divisions: | Faculty of Civil, Environmental, and Geo Engineering > Geomatics Engineering > 29101-(S2) Master Theses |
| Depositing User: | Ilyas Ilyas |
| Date Deposited: | 10 Mar 2025 03:44 |
| Last Modified: | 10 Mar 2025 03:44 |
| URI: | http://repository.its.ac.id/id/eprint/74830 |
Actions (login required)
 |
View Item |