View Item 
      •   IPB Repository
      • Dissertations and Theses
      • Undergraduate Theses
      • UT - Faculty of Agriculture
      • UT - Soil Science and Land Resources
      • View Item
      •   IPB Repository
      • Dissertations and Theses
      • Undergraduate Theses
      • UT - Faculty of Agriculture
      • UT - Soil Science and Land Resources
      • View Item
      JavaScript is disabled for your browser. Some features of this site may not work without it.

      Klasifikasi Tutupan Lahan Rural Menggunakan Citra Sentinel-1 Dengan Metode Support Vector machines (SVM)

      Thumbnail
      View/Open
      Cover (1.127Mb)
      Fulltext (3.105Mb)
      Lampiran (946.0Kb)
      Date
      2023
      Author
      Fadlan, Ahmad
      Trisasongko, Bambang Hendro
      Tjahjono, Boedi
      Metadata
      Show full item record
      Abstract
      Pemetaan klasifikasi tutupan lahan di wilayah rural atau pedesaan saat ini memiliki nilai penting dalam pembangunan berkelanjutan dan pembuatan kebijakan untuk optimalisasi sumber daya alam. Penggunaan Citra radar Sentinel 1 beroperasi optimal dalam semua kondisi cuaca dan mengurangi dampak dari iklim tropis Indonesia yang sering tertutup awan saat perekaman. Penelitian ini bertujuan untuk menelaah karakteristik data hamburan balik dan fitur dekomposisi polarimetrik pada berbagai tutupan lahan rural dan melakukan pemodelan dengan metode Support Vector Machines (SVM). Penelitian dilakukan pada wilayah perbatasan Kabupaten Pandeglang dan Lebak. Pemodelan dilakukan dengan mencoba berbagai kernel dan penyetelan parameter Cost untuk mendapatkan hasil yang optimal. Kinerja terbaik diperoleh dengan penggunaan kernel Radial Basis Function (RBF) dengan akurasi total sebesar 77,74 %. Penyetelan parameter dengan penyesuaian nilai Cost tidak memberikan perubahan yang signifikan pada setiap kernel yang digunakan akibat kompleksitas tutupan lahan yang tinggi. Penggunaan polarimetri parsial dan dekomposisi Eigenvalue belum sepenuhnya efektif dalam mengklasifikasikan tutupan lahan di wilayah yang kompleks.
       
      Land cover classification mapping in rural areas is currently of great importance in sustainable development and policy making for the optimization of natural resources. Sentinel-1 radar imagery operates optimally in all weather conditions and reduces the impact of Indonesia's tropical climate which is often covered by clouds during data acquisition. This study aimed to examine characteristics of backscatter data and polarimetric decomposition features over various rural land covers and to model them using Support Vector Machines (SVM) method. The research was conducted at the border area of Pandeglang and Lebak regencies. Modeling was done by trying various kernels and tuning Cost parameters to get optimal results. The best performance was obtained by using the Radial Basis Function (RBF) kernel with a total accuracy of 77.74%. Parameter tuning by adjusting the Cost value did not provide significant changes in each kernel used due to the high complexity of land cover. The use of partial polarimetry and Eigenvalue decomposition were not fully effective in classifying land cover in complex areas.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/125539
      Collections
      • UT - Soil Science and Land Resources [2825]

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository
        

       

      Browse

      All of IPB RepositoryCollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

      My Account

      Login

      Application

      google store

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository