PERBANDINGAN METODE DENSITY SLICING DAN KLASIFIKASI MULTISPEKTRAL BERBASIS CITRA ALOSAVNIR-2 UNTUK PEMETAAN HUTAN MANGROVE DI ESTUARI SUNGAI ANDAI, MANOKWARI PAPUA BARAT

CHRISTIAN SOLEMAN IMBURI, Drs. Projo Danoedoro, M.Sc, Ph.D; Dr. Sigit Heru Murti B.S., M.Si.

2014 | Tesis | S2 Penginderaan Jauh

Secara umum julat nilai spektral objek diperoleh melalui perpustakaan spektral (spectral library), atau melalui pengukuran langsung di lapangan dengan alat ukur spectroradiometer, spectrometer, dan spectrophotometer, atau alat-alat sejenis yang mahal serta jarang, sehingga sulit untuk memperoleh julat nilai spektral objek tertentu. Studi tentang perbandingan metode density slicing dan klasifikasi multispektral berbasis citra ALOS AVNIR-2 untuk pemetaan hutan mangrove di estuari sungai andai, Manokwari Papua Barat dilakukan dengan tujuan melakukan prosedur analisis perolehan julat nilai spektral melalui analis histogram band inframerah dekat ALOS AVNIR-2. Penelitian ini dilakukan melalui dua tahapan metode analisis, yakni: 1) Analisis utama untuk menghasilkan julat nilai spektral, yakni analisis histogram band inframerah dekat ALOS AVNIR-2, dan 2) Analisis tambahan untuk menjelaskan tingkat akurasi julat nilai spektral melalui: (a) survei lapangan yang menghasilkan data untuk proses uji akurasi, reklasifikasi metode density slicing, dan untuk eksekusi metode klasifikasi terselia, serta (b) melakukan uji akurasi terhadap peta zonasi mangrove berbasis density slicing dan berbasis metode klasifikasi terselia. Kedua analisis tersebut dijabarkan melalui pendekatan penginderaan jauh multitingkat, dimana citra QuickBird digunakan sebagai referensi untuk masking citra ALOS AVNIR-2, dan juga untuk validasi peta hasil perbandingan metode density slicing dan metode klasifikasi terselia. Hasil dari penelitian ini adalah prosedur analisis perolehan julat nilai spektral dan julat nilai spektral zonasi genus mangrove. Selain itu, akurasi keseluruhan untuk metode density slicing 40%, paralellepiped 35,29%, minimum distance 41,18%, mahalanobis distance 58,82%, maximum likelihood 50%, dan spectral angle mapper 58,82%. Penelitian ini secara keseluruhan dapat menghasilkan julat nilai spektral, dan hal ini memberikan implikasi terhadap pengembangan prosedur untuk menghasilkan julat nilai spektral objek.

In general, the range of spectral value objects obtained through spectral library, or through direct measurements in the field with measuring devices spectroradiometer, spectrometer, and spectrophotometer, or with similar tools that are expensive and rare, making it difficult to obtain the range of spectral value for certain objects. The study of A Comparison of Density Slicing Method and Multispectral Classification Method on the Basis of ALOS AVNIR-2 Imagery for Mangrove Mapping in Estuary of Andai River, Manokwari Papua Barat. The purpose of this research is to conduct the analysis procedure acquisition in the range of spectral values through the histogram analysis of near-infrared band of ALOS AVNIR-2. This research was conducted through two stages of analysis methods, namely: 1) the primary analysis to produce the range of spectral values through the histogram analysis of near-infrared band of ALOS AVNIR-2, and 2) additional analysis to explain the accuracy level of the range of spectral values through: (a) field survey that generates the data for the accuracy test process, reclassification of density slicing method, and for execution of supervised classification method, and (b) to test the accuracy of the mangrove zoning map on the basis of density slicing and on the basis of the supervised classification method. Both analyses are elaborated through multilevel remote sensing approach, where the QuickBird imagery is used as a reference for masking ALOS AVNIR-2, and also to validate the map of comparative results of density slicing method and the supervised classification method. The result of this research is the analysis procedure of acquisition range of spectral values and the range of spectral values of mangrove genus zonation. In addition, the overall accuracy for density slicing method is 40%, paralellepiped is 35,29%, 41.18% is the minimum distance, Mahalanobis distance is 58,82%, maximum likelihood is 50%, and spectral angle mapper is 58,82%. The overall study can produce range of spectral values, and this case give implications to the development of procedures in order to generate the range of spectral values of the object.

Kata Kunci : Julat nilai spektral objek, analisis histogram, density slicing, citra ALOS AVNIR-2, citra QuickBird, dan penginderaan jauh multtitingkat