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.

      Pendugaan Leaf Area Index (LAI) Padi (Oryza sativa L.) Berbasis Sentinel-2 Menggunakan Model Deep Learning (1D-CNN)

      Thumbnail
      View/Open
      Cover (425.4Kb)
      Fulltext (2.788Mb)
      Lampiran (567.3Kb)
      Date
      2026
      Author
      MUBAROK, MUHAMMAD NAFIS
      Seminar, Kudang Boro
      Metadata
      Show full item record
      Abstract
      Pendugaan Leaf Area Index (LAI) penting untuk memantau pertumbuhan padi, tetapi pengukuran langsung memerlukan waktu dan tenaga besar sehingga citra Sentinel-2 menjadi alternatif pemantauan wilayah luas. Penelitian ini menganalisis hubungan indeks vegetasi konvensional dan red-edge terhadap LAI, membandingkan model 1D-CNN dengan Random Forest, serta memetakan sebaran LAI. Model dikembangkan untuk fase hingga puncak LAI sebelum pengisian bulir (7–47 hari setelah tanam) menggunakan 273 dari 352 observasi dan 15 fitur hasil seleksi korelasi Pearson dan 5-fold cross-validation. Regresi tersegmentasi menunjukkan breakpoint pada LAI 1,97–3,50 untuk indeks konvensional dan 2,98– 4,39 untuk indeks red-edge, dengan IRECI dan CIred-edge mempertahankan sensitivitas terbesar pada LAI tinggi. Model 1D-CNN hasil Random Search menghasilkan R² uji 0,8245, RMSE 0,9174, dan R² cross-validation 0,7854 ± 0,0381, lebih baik daripada Random Forest yang dioptimasi setara (R² 0,7546). Fitur red-edge menempati empat dari enam peringkat teratas permutation feature importance, sedangkan memasukkan data fase generatif menurunkan R² uji menjadi 0,6875. Validasi eksternal menghasilkan R² negatif dan underprediction pada LAI tinggi sehingga kemampuan transfer model masih terbatas. Peta sebaran LAI dapat mendukung pemantauan pertanaman padi.
       
      Leaf Area Index (LAI) is key to monitoring rice growth, but direct measurement is laborious, so Sentinel-2 offers a large-area alternative. This study analyzed how conventional and red-edge vegetation indices relate to LAI, compared One-Dimensional Convolutional Neural Network with Random Forest, and mapped LAI. The model covered growth up to peak LAI before grain filling (7–47 days after transplanting), using 273 of 352 observations and 15 features selected by Pearson correlation and five-fold cross-validation. Segmented regression showed breakpoints at LAI 1.97–3.50 for conventional indices and 2.98– 4.39 for red-edge indices, with IRECI and CIred-edge retaining the greatest sensitivity at high LAI. The 1D-CNN tuned by Random Search achieved a test R² of 0.8245, RMSE of 0.9174, and cross-validation R² of 0.7854 ± 0.0381, outperforming an equivalently tuned Random Forest (R² 0.7546). Red-edge features occupied four of the top six permutation importance ranks, while including generative-stage data reduced the test R² to 0.6875. External validation produced negative R² and underprediction at high LAI, indicating limited transferability. The LAI maps can support rice crop monitoring.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/180294
      Collections
      • UF - Agricultural and Biosystem Engineering [3666]

      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