Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/157615
Title: Serangan Penyakit Busuk Pangkal Batang Kelapa Sawit (Elaeis guineensis Jacq.) Menggunakan Sentinel-2 di PT Sari Aditya Loka 1
Other Titles: Study of Stem Base Rot Disease Attack of Oil Palm (Elaeis guineensis Jacq.) Using Sentinel-2 at PT Sari Aditya Loka I
Authors: Rochmah, Hidayati Fatchur
Manijo
Hawari, Muhamad Zahran
Issue Date: 2024
Publisher: IPB University
Abstract: Kelapa sawit adalah sektor perkebunan unggulan di Indonesia. Pemantauan perkebunan skala besar memerlukan metode khusus untuk pengumpulan dan analisis data. Pengamatan luas perkebunan telah lama menggunakan data penginderaan jauh. Penelitian ini fokus pada perkiraan nutrisi daun dan identifikasi dampak penyakit tanaman, terutama penyakit busuk pangkal batang (BPB) yang disebabkan oleh cendawan Ganoderma boninense. Data multispektral Sentinel-2 yang tersedia, dimanfaatkan untuk mengidentifikasi gejala awal BPB. Analisis dilakukan dengan menggunakan pendekatan machine learning, yaitu random forest (RF), peran indeks vegetasi, perbandingan produksi, perbandingan Sentinel dan GPS. Hasil model RF yang sudah disetel, diperoleh akurasi di blok OF 02 mencapai 93,90% dan di blok OF 05 mencapai 92,87%, menggunakan lima target berbeda: empat tingkat keparahan dan pohon sehat. Penelitian ini juga menyimpulkan bahwa data tambahan seperti indeks vegetasi dapat membedakan tingkat keparahan penyakit meskipun tidak sepenuhnya responsif. Penggunaan GPS lebih akurat dalam input data dibanding dengan sentinel jika skala kecil. Dampak G. boninense pada blok terserang memiliki penurunan produktivitas kelapa sawit.
Oil palm is the leading plantation sector in Indonesia. Monitoring large-scale plantations requires specialized methods for data collection and analysis. Plantation area observations have long used remote sensing data. This research focuses on estimating leaf nutrition and identifying the impact of plant diseases, especially stem base rot (SBR) caused by the fungus Ganoderma boninense. Available Sentinel-2 multispectral data were utilized to identify early symptoms of SBR. The analysis was conducted using a machine learning approach, namely random forest (RF), the role of vegetation index, production comparison, Sentinel and GPS comparison. The results of the tuned RF model obtained an accuracy of 93.90% in OF 02 block and 92.87% in OF 05 block, using five different targets: four severity levels and healthy trees. The study also concluded that additional data such as vegetation index can differentiate disease severity although it is not fully responsive. Impact of G. boninense on infested blocks has reduced oil palm productivity.
URI: http://repository.ipb.ac.id/handle/123456789/157615
Appears in Collections:UT - Technology and Management of Plantation Production

Files in This Item:
File Description SizeFormat 
cover_J0316201009_f03bf3894b8341f589513e4f14d1b014.pdfCover2.52 MBAdobe PDFView/Open
fulltext_J0316201009_1382d6be17e64daf933a58392448922c.pdf
  Restricted Access
Fulltext8.69 MBAdobe PDFView/Open
lampiran_J0316201009_d7b1c77d266c481bbb011746d87edfd6.pdf
  Restricted Access
Lampiran1.49 MBAdobe PDFView/Open


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