View Item 
      •   IPB Repository
      • Dissertations and Theses
      • Undergraduate Theses
      • UT - Faculty of Mathematics and Natural Sciences
      • UT - Statistics and Data Sciences
      • View Item
      •   IPB Repository
      • Dissertations and Theses
      • Undergraduate Theses
      • UT - Faculty of Mathematics and Natural Sciences
      • UT - Statistics and Data Sciences
      • View Item
      JavaScript is disabled for your browser. Some features of this site may not work without it.

      Kajian ukuran training sample optimum untuk klasifikasi penutup lahan dengan metode kemungkinan maksimum

      Thumbnail
      View/Open
      Full Text (497.1Kb)
      G14amu.pdf (497.1Kb)
      Date
      2014
      Author
      Muhammad, Alfiansyah
      Alamudi, Aam
      Wijayanto, Hari
      Wiweka
      Metadata
      Show full item record
      Abstract
      Klasifikasi dalam penginderaan jauh didefinisikan sebagai suatu metode untuk memberikan label pada piksel berdasarkan karakter spektral yang dimiliki oleh piksel tersebut. Klasifikasi terbimbing merupakan klasifikasi nilai piksel didasarkan pada contoh daerah yang diketahui jenis objek dan nilai spektralnya. Keragaman nilai spektral yang tinggi dan adanya tumpang tindih pada objek yang diamati menjadi permasalahan dalam klasifikasi penutup lahan. Pengambilan training sample yang tepat dan sesuai menjadi sulit dilakukan sehingga mengurangi hasil ketelitian klasifikasi. Kelas penutup lahan diklasifikasikan dengan menggunakan metode kemungkinan maksimum. Kombinasi training sample terbaik adalah kombinasi 9 (111-160 piksel, 15 poligon) dengan rata-rata ketelitian klasifikasi secara keseluruhan (KH) sebesar 70.55% serta rata-rata nilai koefisien Kappa sebesar 0.6298. Training sample perlu dibuat sedemikian rupa sehingga memiliki total piksel besar yang didapatkan dari jumlah poligon besar dengan ukuran masing-masing poligon yang kecil.
       
      Classification in remote sensing is defined as a method for labeling the pixels based on spectral character possessed by the pixel. Supervised classification is a classification of the pixel values are based on sample areas of known object types and spectral values. High spectral value diversity and overlapping on the object observed are problem in land cover classification. Making proper and appropriate training sample becomes difficult to do. It reduces the accuracy of classification results. Land cover classes were classified by maximum likelihood method. The best combination is combination 9 (111-160 pixels, 15 polygons). Average of overall classification accuracy (KH) is 70.55% and the average of Kappa coefficient is 0.6298. Training samples need to be made such that it has a large total pixels obtained from a large number of polygons with a small size of each polygon. Keywords: maximum likelihood classification, remote sensing, training sample
       
      URI
      http://repository.ipb.ac.id/handle/123456789/71780
      Collections
      • UT - Statistics and Data Sciences [2260]

      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