Show simple item record

dc.contributor.advisorJaya, I Nengah Surati
dc.contributor.advisorIlham, Qori Pebrial
dc.contributor.authorOtoluwa, Imam Hawari
dc.date.accessioned2026-08-06T07:39:29Z
dc.date.available2026-08-06T07:39:29Z
dc.date.issued2026
dc.identifier.urihttp://repository.ipb.ac.id/handle/123456789/177437
dc.description.abstractPenelitian ini mengembangkan algoritma untuk mendeteksi kondisi kesehatan hutan mangrove di Pulau Bintan dengan menggunakan pendekatan non-parametrik. Penelitian ini berfokus pada kombinasi beberapa indeks yang dibangun menggunakan data Sentinel-2A dan peubah sosio-geo-biofisik, seperti proximity sungai dan garis pantai, proximity jalan, proximity permukiman, ketinggian, kelerengan, dan substrat. Penelitian ini menunjukkan bahwa model terbaik diperoleh dengan menggunakan kriteria Information Gain dan memberikan akurasi keseluruhan sebesar 95,8% serta akurasi kappa sebesar 95%. SAVI dan Substrat menjadi peubah yang paling berpengaruh dalam memisahkan tutupan lahan, termasuk kelas kesehatan hutan mangrove. Algoritma tersebut berhasil memisahkan kelas kesehatan hutan mangrove dari kelas tutupan lahan lainnya dengan precision berkisar antara 95,6% hingga 98,2% dan recall antara 96,4% hingga 98,2%, serta menunjukkan keefektifan kombinasi peubah spektral dan sosio-geo-biofisik dalam mendeteksi kesehatan hutan mangrove menggunakan data penginderaan jauh di Pulau Bintan.
dc.description.abstractThis research aimed to develop an algorithm for detecting mangrove health conditions on Bintan Island by using a non-parametric approach. The study was focused on examining the combination of several indices derived from Sentinel-2A and socio-geo-biophysical factors such as, proximity to river and coastal line, road distance, proximity to village, elevation, slope, and substrate content. The study showed that the best model was obtained using the Information Gain criterion and provided an overall accuracy of 95.8% and a kappa accuracy of 95%. SAVI and Substrate were identified as the most influential variables in separating land covers which included the mangrove health classes. The algorithm successfully separated mangrove health classes from other land-cover classes having precision ranging from 95.6% to 98.2% and recall between 96.4% to 98.2%, and demonstrated the effectiveness of integrating spectral and socio-geo-biophysical variables for mangrove health detection using remote sensing data on Bintan Island.
dc.description.sponsorship
dc.language.isoid
dc.publisherIPB Universityid
dc.titlePengembangan Algoritma Deteksi Kesehatan Mangrove Berbasis Penginderaan Jauh dan Machine Learning di Pulau Bintanid
dc.title.alternative
dc.typeSkripsi
dc.subject.keywordKesehatan mangroveid
dc.subject.keywordpenginderaan jauhid
dc.subject.keywordpohon keputusanid
dc.subtypeUndergraduate Theses


Files in this item

Thumbnail
Thumbnail
Thumbnail

This item appears in the following Collection(s)

Show simple item record