Mapping of the distribution of oil palm using Landsat 8 data by comparing machine learning and non-machine learning algorithms

N. S.N. Shaharum, H. Z.M. Shafri, W. A.W.A. Ghani, S. Samsatli

Research output: Contribution to journalArticle

Abstract

Oil palm is one of the major crops in Malaysia; it accounts for 47% of the global palm oil supply. Equatorial climate has provided Malaysia with the potential to produce oil palm biomass, which is one of the major contributors to the local economy. The utilisation of oil palm biomass as a source of renewable energy is one of the effective methods to promote green energy. Therefore, there is a need to have sufficient data related to oil palm biomass such as yield estimation, oil palm distributions, and locations. The aim of this study was to produce a land cover map on the distribution of oil palm plantations on three districts located in Selangor. Landsat 8 images of resolutions 15 x 15 m were used and classified via machine learning and non-machine learning algorithms. In this study, three different classifier algorithms were compared using support vector machines, artificial neural networks, and maximum likelihood classifications in which the values obtained for overall accuracy were 98.96%, 99.39%, and 15.30% respectively. The output showed that machine learning algorithms, support vector machines and artificial neural networks gave rise to high accuracies. Hence, the mapping of oil palm distributions via machine learning algorithm was better than that via non-machine learning algorithm.

Original languageEnglish
Pages (from-to)123-135
Number of pages13
JournalPertanika Journal of Tropical Agricultural Science
Volume27
Issue numberS1
Publication statusPublished - 21 Jun 2019
Event4th International Conference on Agricultural & Food Engineering (CAFEi2018) - Kuala Lumpur, Malaysia
Duration: 7 Nov 20189 Nov 2018

Cite this

Mapping of the distribution of oil palm using Landsat 8 data by comparing machine learning and non-machine learning algorithms. / Shaharum, N. S.N.; Shafri, H. Z.M.; Ghani, W. A.W.A.; Samsatli, S.

In: Pertanika Journal of Tropical Agricultural Science, Vol. 27, No. S1, 21.06.2019, p. 123-135.

Research output: Contribution to journalArticle

@article{c032add0c57d40a3aee5855ebf8110f9,
title = "Mapping of the distribution of oil palm using Landsat 8 data by comparing machine learning and non-machine learning algorithms",
abstract = "Oil palm is one of the major crops in Malaysia; it accounts for 47{\%} of the global palm oil supply. Equatorial climate has provided Malaysia with the potential to produce oil palm biomass, which is one of the major contributors to the local economy. The utilisation of oil palm biomass as a source of renewable energy is one of the effective methods to promote green energy. Therefore, there is a need to have sufficient data related to oil palm biomass such as yield estimation, oil palm distributions, and locations. The aim of this study was to produce a land cover map on the distribution of oil palm plantations on three districts located in Selangor. Landsat 8 images of resolutions 15 x 15 m were used and classified via machine learning and non-machine learning algorithms. In this study, three different classifier algorithms were compared using support vector machines, artificial neural networks, and maximum likelihood classifications in which the values obtained for overall accuracy were 98.96{\%}, 99.39{\%}, and 15.30{\%} respectively. The output showed that machine learning algorithms, support vector machines and artificial neural networks gave rise to high accuracies. Hence, the mapping of oil palm distributions via machine learning algorithm was better than that via non-machine learning algorithm.",
author = "Shaharum, {N. S.N.} and Shafri, {H. Z.M.} and Ghani, {W. A.W.A.} and S. Samsatli",
year = "2019",
month = "6",
day = "21",
language = "English",
volume = "27",
pages = "123--135",
journal = "Pertanika Journal of Tropical Agricultural Science",
issn = "1511-3701",
publisher = "Universiti Putra Malaysia",
number = "S1",

}

TY - JOUR

T1 - Mapping of the distribution of oil palm using Landsat 8 data by comparing machine learning and non-machine learning algorithms

AU - Shaharum, N. S.N.

AU - Shafri, H. Z.M.

AU - Ghani, W. A.W.A.

AU - Samsatli, S.

PY - 2019/6/21

Y1 - 2019/6/21

N2 - Oil palm is one of the major crops in Malaysia; it accounts for 47% of the global palm oil supply. Equatorial climate has provided Malaysia with the potential to produce oil palm biomass, which is one of the major contributors to the local economy. The utilisation of oil palm biomass as a source of renewable energy is one of the effective methods to promote green energy. Therefore, there is a need to have sufficient data related to oil palm biomass such as yield estimation, oil palm distributions, and locations. The aim of this study was to produce a land cover map on the distribution of oil palm plantations on three districts located in Selangor. Landsat 8 images of resolutions 15 x 15 m were used and classified via machine learning and non-machine learning algorithms. In this study, three different classifier algorithms were compared using support vector machines, artificial neural networks, and maximum likelihood classifications in which the values obtained for overall accuracy were 98.96%, 99.39%, and 15.30% respectively. The output showed that machine learning algorithms, support vector machines and artificial neural networks gave rise to high accuracies. Hence, the mapping of oil palm distributions via machine learning algorithm was better than that via non-machine learning algorithm.

AB - Oil palm is one of the major crops in Malaysia; it accounts for 47% of the global palm oil supply. Equatorial climate has provided Malaysia with the potential to produce oil palm biomass, which is one of the major contributors to the local economy. The utilisation of oil palm biomass as a source of renewable energy is one of the effective methods to promote green energy. Therefore, there is a need to have sufficient data related to oil palm biomass such as yield estimation, oil palm distributions, and locations. The aim of this study was to produce a land cover map on the distribution of oil palm plantations on three districts located in Selangor. Landsat 8 images of resolutions 15 x 15 m were used and classified via machine learning and non-machine learning algorithms. In this study, three different classifier algorithms were compared using support vector machines, artificial neural networks, and maximum likelihood classifications in which the values obtained for overall accuracy were 98.96%, 99.39%, and 15.30% respectively. The output showed that machine learning algorithms, support vector machines and artificial neural networks gave rise to high accuracies. Hence, the mapping of oil palm distributions via machine learning algorithm was better than that via non-machine learning algorithm.

M3 - Article

VL - 27

SP - 123

EP - 135

JO - Pertanika Journal of Tropical Agricultural Science

JF - Pertanika Journal of Tropical Agricultural Science

SN - 1511-3701

IS - S1

ER -