IPB University Logo

SCIENTIFIC REPOSITORY

IPB University Scientific Repository collects, disseminates, and provides persistent and reliable access to the research and scholarship of faculty, staff, and students at IPB University

AI Repository
 
Building and Categories


      View Item 
      •   IPB Repository
      • Final Assignments
      • Undergraduate Final Assignments
      • UF - Faculty of Agricultural Technology
      • UF - Agricultural and Biosystem Engineering
      • View Item
      •   IPB Repository
      • Final Assignments
      • Undergraduate Final Assignments
      • UF - Faculty of Agricultural Technology
      • UF - Agricultural and Biosystem Engineering
      • View Item
      JavaScript is disabled for your browser. Some features of this site may not work without it.

      Integrasi Sentinel-1 dengan Algoritma Deep Learning dalam Pendugaan Fase Pertanaman Padi (Oryza sativa.L)

      Thumbnail
      View/Open
      Cover (486.9Kb)
      Fulltext (3.276Mb)
      Lampiran (3.360Mb)
      Date
      2026
      Author
      RIZKILLAH, IRFAN ZIYADI
      Seminar, Kudang Boro
      Metadata
      Show full item record
      Abstract
      Informasi fase pertanaman padi diperlukan untuk memantau kondisi pertanaman secara spasial dan berkala. Penelitian ini bertujuan membangun model klasifikasi enam fase pertanaman padi sawah, yaitu pengairan, vegetatif 1, vegetatif 2, generatif 1, generatif 2, dan bera, berbasis deret waktu Sentinel-1 menggunakan Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), dan hibrida CNN–LSTM, serta membandingkan kinerja ketiganya. Penelitian berlangsung pada Mei–Agustus 2026 di Kecamatan Rawamerta, Kabupaten Karawang, dan Kecamatan Cisaat, Kabupaten Sukabumi. Data pemodelan mencakup 2025 observasi dari 81 titik pada 25 tanggal akuisisi tahun 2020 di Rawamerta. Data dibagi menjadi data pengembangan dan data uji berdasarkan lokasi titik sampel. Pemodelan mencakup seleksi fitur, pembentukan sekuens empat pengamatan, dan hyperparameter tuning melalui validasi silang GroupKFold empat fold dengan tiga seed. Evaluasi pada 425 sampel uji menunjukkan CNN memperoleh kinerja tertinggi dengan akurasi 90,82% dan Macro-F1 score 0,9142. Validasi eksternal membandingkan prediksi terhadap pengamatan lapangan tahun 2026 untuk menilai generalisasi lintas musim dan lintas wilayah. Secara deskriptif, kesesuaian mencapai 95,06% (77 dari 81 titik) di Rawamerta dan 78,95% (15 dari 19 titik) di Cisaat. Kedua persentase tidak ditafsirkan sebagai perbandingan langsung karena ukuran sampel dan karakteristik skenario validasi berbeda. Penerapan model menghasilkan peta distribusi fase pertanaman dan estimasi luas lahan sawah setiap fase untuk mendukung pemantauan pertanaman padi.
       
      Rice crop stage information is needed for spatial and periodic crop monitoring. This study aimed to build models for classifying six lowland rice crop stages, namely flooding, vegetative 1, vegetative 2, generative 1, generative 2, and fallow, from Sentinel-1 time series using a Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and a hybrid CNN–LSTM, and to compare their performance. The study was conducted from May to August 2026 in Rawamerta District, Karawang Regency, and Cisaat District, Sukabumi Regency. Modeling data comprised 2025 observations from 81 points across 25 acquisition dates in Rawamerta in 2020. The data were divided into development and test sets based on sampling point location. Modeling involved feature selection, four-observation sequence construction, and hyperparameter optimization using four-fold GroupKFold cross-validation with three seeds. Evaluation on 425 test samples showed that CNN achieved the highest performance, with 90,82% accuracy and a Macro-F1 score of 0,9142. External validation compared predictions with 2026 field observations to assess generalization across seasons and regions. Descriptively, agreement reached 95,06% (77 of 81 points) in Rawamerta and 78,95% (15 of 19 points) in Cisaat. These percentages were not interpreted as directly comparable because the sample sizes and validation settings differed. Model application produced crop stage distribution maps and paddy field area estimates for each stage to support rice crop monitoring.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/180257
      Collections
      • UF - Agricultural and Biosystem Engineering [3662]

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository
        

       

      Browse

      All of IPB RepositoryCollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

      My Account

      Login

      Application

      google store

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository