Show simple item record

dc.contributor.advisorSeminar, Kudang Boro
dc.contributor.advisorSudradjat
dc.contributor.authorFajrisani, Fadillah
dc.date.accessioned2025-08-10T11:33:17Z
dc.date.available2025-08-10T11:33:17Z
dc.date.issued2025
dc.identifier.urihttp://repository.ipb.ac.id/handle/123456789/168563
dc.description.abstractBesi (Fe) merupakan unsur hara mikro esensial yang berperan penting dalam menunjang fungsi fisiologis dan produktivitas kelapa sawit. Penentuan status hara Fe secara konvensional memiliki sejumlah keterbatasan, seperti bersifat destruktif, membutuhkan waktu dan biaya tinggi, serta kurang mampu merepresentasikan variasi spasial dalam skala luas. Penelitian ini bertujuan mengembangkan model prediksi Fe berbasis citra Sentinel-2A menggunakan algoritma Support Vector Regression untuk estimasi status Fe pada kelapa sawit di lahan gambut, mineral, serta gabungan keduanya. Data yang digunakan mencakup hasil analisis laboratorium kadar Fe pada daun sebagai variabel target, serta citra Sentinel-2A dan indeks vegetasi sebagai variabel input. Pengolahan citra meliputi resampling ke resolusi 10 × 10 meter, penghapusan awan, dan ekstraksi nilai piksel. Dataset dibagi menjadi 80% data latih dan 20% data uji. Hasil menunjukkan bahwa model spesifik lahan memberikan prediksi kandungan Fe yang lebih presisi dibandingkan model gabungan, dengan selisih absolut yang lebih rendah sebesar 8,39 ppm (gambut) dan 22,50 ppm (mineral), dibandingkan dengan model gabungan sebesar 10,64 ppm (gambut) dan 25,75 ppm (mineral). Hal ini menunjukkan bahwa pendekatan spesifik lebih unggul untuk estimasi Fe dan mendukung rekomendasi pemupukan yang presisi. Secara umum, model mampu memetakan variabilitas spasial Fe dengan baik serta merepresentasikan perbedaan karakteristik antara lahan gambut dan mineral.
dc.description.abstractIron (Fe) is an essential micronutrient that plays an important role in supporting the physiological functions and productivity of oil palms. Conventional methods for determining Fe nutrient status have a number of limitations, such as being destructive, time-consuming, costly, and unable to represent spatial variations on a large scale. This study aims to develop an iron prediction model based on Sentinel-2A imagery using the Support Vector Regression algorithm to estimate iron status in oil palms on peatland, mineral soil, and a combination of both. The data used includes laboratory analysis results of iron content in leaves as the target variable, as well as Sentinel-2A imagery and vegetation indices as input variables. Image processing includes resampling to a resolution of 10 × 10 meters, cloud removal, and pixel value extraction. The dataset is divided into 80% training data and 20% test data. The results show that the land-specific model provides more precise Fe content predictions than the combined model, with lower absolute differences of 8,39 ppm (peat) and 22,5 ppm (mineral), compared to the combined model of 10,64 ppm (peat) and 25,75 ppm (mineral). This indicates that the specific approach is superior for Fe estimation and supports precise fertilization recommendations. Overall, the model effectively maps the spatial variability of Fe and represents the differences in characteristics between peat and mineral soils.
dc.description.sponsorshipTim Riset PreciPalm (Precision Agriculture Platform for Oil Palm)
dc.language.isoid
dc.publisherIPB Universityid
dc.titleEstimasi Status Hara Besi (Fe) pada Tanaman Kelapa Sawit Berbasis Citra Sentinel-2Aid
dc.title.alternativeEstimation of Iron (Fe) Nutrient Status in Oil Palm Plants Based on Sentinel-2A Imagery
dc.typeSkripsi
dc.subject.keywordironid
dc.subject.keywordoil palmid
dc.subject.keywordsentinel-2Aid
dc.subject.keywordvegetation indexid
dc.subject.keywordsupport vector regressionid


Files in this item

Thumbnail
Thumbnail
Thumbnail

This item appears in the following Collection(s)

Show simple item record