Synthetic Minority Over-Sampling Technique (SMOTE) and Logistic Model Tree (LMT)-adaptive boosting algorithms for classifying imbalanced datasets of nutrient and chlorophyll sufficiency levels of oil palm (Elaeis guineensis) using spectroradiometers and unmanned aerial vehicles
The conventional method to quantify leaf biochemical properties (nutrients and chlorophylls) is tedious, labour-intensive, and impractical for vast oil palm plantation areas. Spectral analysis retrieved from a spectroradiometer and an unmanned aerial vehicle (UAV) and imbalanced approaches such as t...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Published: |
Elsevier
2022
|
| _version_ | 1825938847838502912 |
|---|---|
| author | Amirruddin, Amiratul Diyana Muharam, Farrah Melissa Ismail, Mohd Hasmadi Tan, Ngai Paing Ismail, Mohd Firdaus |
| author_facet | Amirruddin, Amiratul Diyana Muharam, Farrah Melissa Ismail, Mohd Hasmadi Tan, Ngai Paing Ismail, Mohd Firdaus |
| author_sort | Amirruddin, Amiratul Diyana |
| collection | UPM |
| description | The conventional method to quantify leaf biochemical properties (nutrients and chlorophylls) is tedious, labour-intensive, and impractical for vast oil palm plantation areas. Spectral analysis retrieved from a spectroradiometer and an unmanned aerial vehicle (UAV) and imbalanced approaches such as the Synthetic Minority Over-sampling TEchnique (SMOTE) and machine learning have given promising results for monitoring plant biochemical properties. However, the integration of these methods is not widely explored for oil palm. There are three primary aims of the current study. We evaluate the effectiveness of the integration of SMOTE, Logistic Model Tree (LMT), and Adaptive Boosting (AdaBoost) to address data imbalance problems for the assessment of the oil palm nutrients and chlorophylls status. The performance of the raw band and vegetation index (VI) extracted from the UAV in assessing leaf biochemical properties of mature oil palms is also addressed. Finally, we compare the competency of the spectral model retrieved from the spectroradiometer and UAV. In the study, nitrogen (N) treatments varying between 0 and 6 kg palm−1 were applied to mature Tenera palms. The integration of SMOTE with LMT and AdaBoost (LMT-SMOTEBoost) outperformed other approaches in classifying the leaf biochemical sufficiency status of mature oil palm. The VIs outperformed the raw band in discriminating the leaf biochemical properties at the canopy level. Both leaf and canopy spectral models obtained from spectroradiometer and UAV were comparable and produced good performance with balanced accuracy (BAcc) above 0.77. Using these techniques may provide palm oil plantation owners with a cost-effective way to monitor nutrient levels in palms more efficiently and comprehensively to ensure greater harvests and tree health. |
| first_indexed | 2024-03-06T11:18:24Z |
| format | Article |
| id | upm.eprints-103414 |
| institution | Universiti Putra Malaysia |
| last_indexed | 2024-03-06T11:18:24Z |
| publishDate | 2022 |
| publisher | Elsevier |
| record_format | dspace |
| spelling | upm.eprints-1034142023-06-13T07:11:34Z http://psasir.upm.edu.my/id/eprint/103414/ Synthetic Minority Over-Sampling Technique (SMOTE) and Logistic Model Tree (LMT)-adaptive boosting algorithms for classifying imbalanced datasets of nutrient and chlorophyll sufficiency levels of oil palm (Elaeis guineensis) using spectroradiometers and unmanned aerial vehicles Amirruddin, Amiratul Diyana Muharam, Farrah Melissa Ismail, Mohd Hasmadi Tan, Ngai Paing Ismail, Mohd Firdaus The conventional method to quantify leaf biochemical properties (nutrients and chlorophylls) is tedious, labour-intensive, and impractical for vast oil palm plantation areas. Spectral analysis retrieved from a spectroradiometer and an unmanned aerial vehicle (UAV) and imbalanced approaches such as the Synthetic Minority Over-sampling TEchnique (SMOTE) and machine learning have given promising results for monitoring plant biochemical properties. However, the integration of these methods is not widely explored for oil palm. There are three primary aims of the current study. We evaluate the effectiveness of the integration of SMOTE, Logistic Model Tree (LMT), and Adaptive Boosting (AdaBoost) to address data imbalance problems for the assessment of the oil palm nutrients and chlorophylls status. The performance of the raw band and vegetation index (VI) extracted from the UAV in assessing leaf biochemical properties of mature oil palms is also addressed. Finally, we compare the competency of the spectral model retrieved from the spectroradiometer and UAV. In the study, nitrogen (N) treatments varying between 0 and 6 kg palm−1 were applied to mature Tenera palms. The integration of SMOTE with LMT and AdaBoost (LMT-SMOTEBoost) outperformed other approaches in classifying the leaf biochemical sufficiency status of mature oil palm. The VIs outperformed the raw band in discriminating the leaf biochemical properties at the canopy level. Both leaf and canopy spectral models obtained from spectroradiometer and UAV were comparable and produced good performance with balanced accuracy (BAcc) above 0.77. Using these techniques may provide palm oil plantation owners with a cost-effective way to monitor nutrient levels in palms more efficiently and comprehensively to ensure greater harvests and tree health. Elsevier 2022 Article PeerReviewed Amirruddin, Amiratul Diyana and Muharam, Farrah Melissa and Ismail, Mohd Hasmadi and Tan, Ngai Paing and Ismail, Mohd Firdaus (2022) Synthetic Minority Over-Sampling Technique (SMOTE) and Logistic Model Tree (LMT)-adaptive boosting algorithms for classifying imbalanced datasets of nutrient and chlorophyll sufficiency levels of oil palm (Elaeis guineensis) using spectroradiometers and unmanned aerial vehicles. Computers and Electronics in Agriculture, 193. art. no. 106646. pp. 1-16. ISSN 0168-1699; ESSN: 1872-7107 https://www.sciencedirect.com/science/article/pii/S0168169921006633 10.1016/j.compag.2021.106646 |
| spellingShingle | Amirruddin, Amiratul Diyana Muharam, Farrah Melissa Ismail, Mohd Hasmadi Tan, Ngai Paing Ismail, Mohd Firdaus Synthetic Minority Over-Sampling Technique (SMOTE) and Logistic Model Tree (LMT)-adaptive boosting algorithms for classifying imbalanced datasets of nutrient and chlorophyll sufficiency levels of oil palm (Elaeis guineensis) using spectroradiometers and unmanned aerial vehicles |
| title | Synthetic Minority Over-Sampling Technique (SMOTE) and Logistic Model Tree (LMT)-adaptive boosting algorithms for classifying imbalanced datasets of nutrient and chlorophyll sufficiency levels of oil palm (Elaeis guineensis) using spectroradiometers and unmanned aerial vehicles |
| title_full | Synthetic Minority Over-Sampling Technique (SMOTE) and Logistic Model Tree (LMT)-adaptive boosting algorithms for classifying imbalanced datasets of nutrient and chlorophyll sufficiency levels of oil palm (Elaeis guineensis) using spectroradiometers and unmanned aerial vehicles |
| title_fullStr | Synthetic Minority Over-Sampling Technique (SMOTE) and Logistic Model Tree (LMT)-adaptive boosting algorithms for classifying imbalanced datasets of nutrient and chlorophyll sufficiency levels of oil palm (Elaeis guineensis) using spectroradiometers and unmanned aerial vehicles |
| title_full_unstemmed | Synthetic Minority Over-Sampling Technique (SMOTE) and Logistic Model Tree (LMT)-adaptive boosting algorithms for classifying imbalanced datasets of nutrient and chlorophyll sufficiency levels of oil palm (Elaeis guineensis) using spectroradiometers and unmanned aerial vehicles |
| title_short | Synthetic Minority Over-Sampling Technique (SMOTE) and Logistic Model Tree (LMT)-adaptive boosting algorithms for classifying imbalanced datasets of nutrient and chlorophyll sufficiency levels of oil palm (Elaeis guineensis) using spectroradiometers and unmanned aerial vehicles |
| title_sort | synthetic minority over sampling technique smote and logistic model tree lmt adaptive boosting algorithms for classifying imbalanced datasets of nutrient and chlorophyll sufficiency levels of oil palm elaeis guineensis using spectroradiometers and unmanned aerial vehicles |
| work_keys_str_mv | AT amirruddinamiratuldiyana syntheticminorityoversamplingtechniquesmoteandlogisticmodeltreelmtadaptiveboostingalgorithmsforclassifyingimbalanceddatasetsofnutrientandchlorophyllsufficiencylevelsofoilpalmelaeisguineensisusingspectroradiometersandunmannedaerialvehicles AT muharamfarrahmelissa syntheticminorityoversamplingtechniquesmoteandlogisticmodeltreelmtadaptiveboostingalgorithmsforclassifyingimbalanceddatasetsofnutrientandchlorophyllsufficiencylevelsofoilpalmelaeisguineensisusingspectroradiometersandunmannedaerialvehicles AT ismailmohdhasmadi syntheticminorityoversamplingtechniquesmoteandlogisticmodeltreelmtadaptiveboostingalgorithmsforclassifyingimbalanceddatasetsofnutrientandchlorophyllsufficiencylevelsofoilpalmelaeisguineensisusingspectroradiometersandunmannedaerialvehicles AT tanngaipaing syntheticminorityoversamplingtechniquesmoteandlogisticmodeltreelmtadaptiveboostingalgorithmsforclassifyingimbalanceddatasetsofnutrientandchlorophyllsufficiencylevelsofoilpalmelaeisguineensisusingspectroradiometersandunmannedaerialvehicles AT ismailmohdfirdaus syntheticminorityoversamplingtechniquesmoteandlogisticmodeltreelmtadaptiveboostingalgorithmsforclassifyingimbalanceddatasetsofnutrientandchlorophyllsufficiencylevelsofoilpalmelaeisguineensisusingspectroradiometersandunmannedaerialvehicles |