Klasifikasi Fase Pertumbuhan Padi Berdasarkan Data Band Dan Indeks Vegetasi Citra Multitemporal Landsat-8 Dengan Metode Deep Neural Network (Studi Kasus Sampel Survei Ksa Kabupaten Poso, Provinsi Sulawesi Tengah)

Rifki, Kevin Agung Fernanda (2022) Klasifikasi Fase Pertumbuhan Padi Berdasarkan Data Band Dan Indeks Vegetasi Citra Multitemporal Landsat-8 Dengan Metode Deep Neural Network (Studi Kasus Sampel Survei Ksa Kabupaten Poso, Provinsi Sulawesi Tengah). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06211840000031-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Survei Kerangka Sampel Area (KSA) yang dilaksanakan oleh BPS mulai tahun 2018 memiliki keterbatasan. Maka untuk mengatasi keterbatasan tersebut, maka dilakukan menggunakan citra satelit Landsat-8 dengan metode Deep Neural Network (DNN) untuk mengklasifikasikan fase pertumbuhan tanaman area sawah khususnya padi. Inovasi yang dikembangkan melalui kombinasi data satelit dengan data ofisial untuk memberikan solusi dari keterbatasan Survei KSA yang dilaksanakan oleh BPS. Data juga ditambah dengan indeks vegetasi hasil perhitungan dari data satelit dan dikombinasikan dengan data ofisial. Deep Neural Network merupakan salah satu metode machine learning berbasis jaringan syaraf manusia hasil pengembangan dari Artificial Neural Network (ANN) dimana dibuat beberapa hidden layer sehingga bisa mengklasifikasi data yang cukup kompleks. Penelitian ini bertujuan mendapatkan metode yang optimal antara ANN dan DNN dalam mengklasifikasi fase pertumbuhan padi. Hasil klasifikasi menunjukkan secara rata-rata kinerja dan nilai Average Precision (AP) baik dari ANN dan DNN tidak ada perbedaan yang signifikan. Dengan pembagian data stratified 5-fold cross validation, diperoleh baik antara ANN dan DNN terpilih fold ke-5 dan didapatkan model DNN yang paling optimum baik secara kinerja klasifikasi dari accuracy, precision, sensitivity, fl-score, cohen kappa index, dan nilai AP untuk mengklasifikasi fase pertumbuhan padi. Proses prediksi dilakukan dengan menggunakan model DNN dan dapat memprediksi fase pertumbuhan padi secara tepat pada 26 dari 46 titik sampel amatan namun model tersebut belum bisa memprediksi fase pertumbuhan padi Survei KSA dengan baik.
==============================================================================================================================
The Area Sample Framework Survey (KSA) conducted by BPS starting in 2018 has limitations. So, to overcome these limitations, it was carried out using Landsat-8 satellite imagery with the Deep Neural Network (DNN) method to classify the growth phases of paddy fields, especially rice. An innovation developed through a combination of satellite data with official data to provide a solution to the limitations of the KSA Survey conducted by BPS. Data is also added to the vegetation index calculated from satellite data and combined with official data. Deep Neural Network is one of the machine learning methods based on human neural networks, as a result of the development of the Artificial Neural Network (ANN) where several hidden layers are created so that they can classify fairly complex data. This study aims to obtain the optimal method between ANN and DNN in classifying rice growth phases. The classification results show that there is no significant difference between the performance and Average Precision (AP) of both ANN and DNN. By dividing the stratified 5-fold cross-validation data, it was obtained that both ANN and DNN were selected for the 5th fold and the most optimum DNN model was obtained both in terms of classification performance from accuracy, precision, sensitivity, fl-score, cohen kappa index, and AP value. to classify rice growth phases. The prediction process is carried out using the DNN model and can predict the rice growth phase correctly at 26 of the 46 sample points observed, but the model has not been able to predict the rice growth phase of the ASF survey well.

Item Type:	Thesis (Other)
Additional Information:	RSSt 519.53 Rif k-1 2022
Uncontrolled Keywords:	Deep Neural Network. Kerangka Sampel Area. Klasifikasi. Landsat-8. Area Sampling Framework. Classification. Deep Neural Network. Landsat-8.
Subjects:	H Social Sciences > HA Statistics
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	10 Jun 2026 04:41
Last Modified:	10 Jun 2026 04:41
URI:	http://repository.its.ac.id/id/eprint/133691

Actions (login required)

View Item