Klasifikasi Tanaman Atau Bukan Tanaman Pada Data Satelit Earth Observation Dengan Metode Deep Learning

Pratama, Refaldyka Galuh (2023) Klasifikasi Tanaman Atau Bukan Tanaman Pada Data Satelit Earth Observation Dengan Metode Deep Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 05111940000209-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (4MB) | Request a copy

Abstract

Populasi manusia yang terus menerus bertambah menjadi tantangan bagi dunia untuk menangani permasalahan manusia, yaitu makin bertambahnya kebutuhan manusia dalam hal pangan. Dalam hal ini, pertanian memegang peran penting untuk mengatasi masalah tersebut. Sumber daya pangan seperti nasi, gandum, dan jagung atau bahkan bahan pokok pangan lainnya memegang peranan yang sangat penting bagi pertumbuhan populasi manusia. Maka dari itu, informasi distribusi dan kondisi sumber daya pangan ini dibutuhkan. Dataset CropHarvest telah digunakan dalam penelitian sebelumnya untuk mengklasifikasikan tumbuhan atau bukan tumbuhan. Akan tetapi, model yang digunakan untuk melatih dataset tersebut kurang optimal. Oleh karena itu, pada penelitian ini dataset CropHarvest dilatih dengan menggunakan model Random Forest, Deep Forest dan Convolutional Neural Network (CNN). Setelah itu, model dievaluasi dengan cara membandingkan nilai akurasi, nilai AUC_ROC, nilai F1, nilai precision, dan nilai recall pada setiap model. Berdasarkan hasil evaluasi, pada dataset Kenya performa terbaik pada setiap metriks adalah nilai akurasi 95,27%, nilai AUC_ROC 91,84%, nilai F1 72,08%, nilai precision 69% dan nilai recall 88,75%. Pada dataset Kenya under sample performa terbaik pada setiap metriks adalah nilai akurasi 93,54%, nilai AUC_ROC 95,84%, nilai F1 70,39%, nilai precision 56,31%, dan nilai recall 100%. Kemudian, pada dataset Kenya over sample performa terbaik pada setiap metriks adalah nilai akurasi 94,63%, nilai AUC_ROC 94,39%, nilai F1 73,10%, nilai precision 75,86%, dan nilai recall 90%. Terakhir, pada dataset Togo performa terbaik pada setiap metriks adalah nilai akurasi 89,53%, nilai AUC_ROC 89,30%, dan nilai F1 90,78%, nilai precision 91,24%, dan nilai recall 91,10%.
=================================================================================================================================
The continuously increasing human population will be a challenge for the world to deal with human problems, which is the increasing human need for food. In this case, agriculture plays an important role to overcome the problem. Food resources such as rice, wheat and corn or even other food staples play a very important role for human population growth. Therefore, information on the distribution and condition of these food resources is needed. The CropHarvest dataset has been used in previous research to classify crop vs noncrop. However, the model used to train the dataset was not optimal. Therefore, in this study the CropHarvest dataset will be trained using the Random Forest model, Deep Forest model and Convolutional Neural Network (CNN) model. After that, the model will be evaluated by comparing the accuracy score, AUC_ROC score, F1 score, precision score, and recall score. Based on the evaluation results, the best performance on each metric in the Kenya dataset is as follows: the accuracy is 95.27%, AUC_ROC is 91.84%, F1 score is 72.08%, precision is 69%, and recall is 88.75%. In the under-sampled Kenya dataset, the best performance on each metric is: accuracy is 93.54%, AUC_ROC is 95.84%, F1 score is 70.39%, precision is 56.31%, and recall is 100%. Furthermore, in the over-sampled Kenya dataset, the best performance on each metric is: accuracy is 94.63%, AUC_ROC is 94.39%, F1 score is 73.10%, precision is 75.86%, and recall is 90%. Lastly, in the Togo dataset, the best performance on each metric is: accuracy is 89.53%, AUC_ROC is 89.30%, F1 score is 90.78%, precision is 91.24%, and recall is 91.10%.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	CropHarvest, Deep Learning, Deep Forest, Klasifikasi, Classification
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. R Medicine > R Medicine (General) > R858 Deep Learning T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Refaldyka Galuh Pratama
Date Deposited:	18 Oct 2023 08:41
Last Modified:	18 Oct 2023 08:41
URI:	http://repository.its.ac.id/id/eprint/101985

Actions (login required)

View Item