Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/115744
Title: Klasifikasi Biomassa Atas Permukaan Pada Perkebunan Sawit Menggunakan Metode Random Forest
Other Titles: Aboveground Biomass Classification in Oil Palm Plantation using Random Forest Method
Authors: Barus, Baba
Ardiansyah, Muhammad
Puspita, Gita
Issue Date: Dec-2022
Publisher: IPB University
Citation: Puspita 2022
Abstract: Kelapa sawit merupakan salah satu tumbuhan penyerap karbon terbesar di Indonesia. Karbon disimpan dalam bentuk biomassa dari kelapa sawit dapat dihitung menggunakan metode non-destruktif dengan persamaan alometrik. Pemantauan biomassa terutama pada area perkebunan sawit yang luas dapat digunakan metode penginderaan jauh yang dikombinasikan dengan machine learning seperti klasifikasi random forest (RF). Tujuan dari penelitian ini adalah untuk menduga nilai biomassa atas permukaan di perkebunan kelapa sawit menurut kelas umur menggunakan metode non destruktif, menganalisis hubungan antara reflektansi kelapa sawit yang berasal dari citra Sentinel-2 dan biomassa atas permukaan kelapa sawit, dan memetakan sebaran biomassa atas permukaan kelapa sawit menggunakan metode RF. Hasil penelitian menunjukkan adanya korelasi positif antara biomassa atas permukaan dengan umur kelapa sawit. Nilai reflektan citra Sentinel-2 pada semua band yaitu blue, green, red, dan NIR memiliki korelasi positif dengan biomassa sawit. Klasifikasi biomassa menggunakan RF menghasilkan nilai akurasi keseluruhan sebesar 76,4% pada kombinasi training dan testing data 60% : 40%.
Oil palm is one of the largest carbon absorbing plants in Indonesia. Carbon stored in the form of biomass from oil palm can be quantified using non-destructive methods with allometric equations. To monitor biomass, especially in large areas of oil palm plantations, remote sensing methods can be used combined with machine learning such as random forest (RF) classification. The purpose of this study was to estimate the value of aboveground biomass in oil palm plantations according to age class using non-destructive methods, to analyze the relationship between the reflectance of oil palm derived from Sentinel-2 imagery and the aboveground biomass of oil palm, and to map the distribution of aboveground biomass of oil palm using the RF method. The results showed a positive correlation between the aboveground biomass and the age of the oil palm. The reflectance value of Sentinel-2 imagery in all bands namely Blue, Green, Red, and NIR has a positive correlation with oil palm biomass. Classification of biomass using RF produces an overall accuracy value of 76.4% on a combination of training and testing data 60%: 40%.
URI: http://repository.ipb.ac.id/handle/123456789/115744
Appears in Collections:UT - Soil Science and Land Resources

Files in This Item:
File Description SizeFormat 
COVER_GITA PUSPITA (A14170022).pdf
  Restricted Access
Cover464.69 kBAdobe PDFView/Open
A14170022_Gita Puspita.pdf
  Restricted Access
Fullteks1.78 MBAdobe PDFView/Open
LAMPIRAN_Gita Puspita.pdf
  Restricted Access
Lampiran1.78 MBAdobe PDFView/Open


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