MHRA-MS-3D-ResNet-BiLSTM: A Multi-Head-Residual Attention-Based Multi-Stream Deep Learning Model for Soybean Yield Prediction in the U.S. Using Multi-Source Remote Sensing Data

Publikation: Beitrag in FachzeitschriftArtikelForschungPeer-Review

Autorschaft

Externe Organisationen

  • University of Tehran
  • Hochschule für Wirtschaft und Gesellschaft Ludwigshafen
Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Aufsatznummer107
Seitenumfang25
FachzeitschriftRemote sensing
Jahrgang17
Ausgabenummer1
PublikationsstatusVeröffentlicht - 31 Dez. 2024

Abstract

Accurate prediction of soybean yield is important for safeguarding food security and improving agricultural management. Recent advances have highlighted the effectiveness and ability of Machine Learning (ML) models in analyzing Remote Sensing (RS) data for this purpose. However, most of these models do not fully consider multi-source RS data for prediction, as processing these increases complexity and limits their accuracy and generalizability. In this study, we propose the Multi-Residual Attention-Based Multi-Stream 3D-ResNet-BiLSTM (MHRA-MS-3D-ResNet-BiLSTM) model, designed to integrate various RS data types, including Sentinel-1/2 imagery, Daymet climate data, and soil grid information, for improved county-level U.S. soybean yield prediction. Our model employs a multi-stream architecture to process diverse data types concurrently, capturing complex spatio-temporal features effectively. The 3D-ResNet component utilizes 3D convolutions and residual connections for pattern recognition, complemented by Bidirectional Long Short-Term Memory (BiLSTM) for enhanced long-term dependency learning by processing data arrangements in forward and backward directions. An attention mechanism further refines the model’s focus by dynamically weighting the significance of different input features for efficient yield prediction. We trained the MHRA-MS-3D-ResNet-BiLSTM model using multi-source RS datasets from 2019 and 2020 and evaluated its performance with U.S. soybean yield data for 2021 and 2022. The results demonstrated the model’s robustness and adaptability to unseen data, achieving an R2 of 0.82 and a Mean Absolute Percentage Error (MAPE) of 9% in 2021, and an R2 of 0.72 and MAPE of 12% in 2022. This performance surpassed some of the state-of-the-art models like 3D-ResNet-BiLSTM and MS-3D-ResNet-BiLSTM, and other traditional ML methods like Random Forest (RF), XGBoost, and LightGBM. These findings highlight the methodology’s capability to handle multiple RS data types and its role in improving yield predictions, which can be helpful for sustainable agriculture.

ASJC Scopus Sachgebiete

Ziele für nachhaltige Entwicklung

Zitieren

MHRA-MS-3D-ResNet-BiLSTM: A Multi-Head-Residual Attention-Based Multi-Stream Deep Learning Model for Soybean Yield Prediction in the U.S. Using Multi-Source Remote Sensing Data. / Fathi, Mahdiyeh; Shah-Hosseini, Reza; Moghimi, Armin et al.
in: Remote sensing, Jahrgang 17, Nr. 1, 107, 31.12.2024.

Publikation: Beitrag in FachzeitschriftArtikelForschungPeer-Review

Download
@article{7ab24d95aeb145b89b2853be58eb4059,
title = "MHRA-MS-3D-ResNet-BiLSTM: A Multi-Head-Residual Attention-Based Multi-Stream Deep Learning Model for Soybean Yield Prediction in the U.S. Using Multi-Source Remote Sensing Data",
abstract = "Accurate prediction of soybean yield is important for safeguarding food security and improving agricultural management. Recent advances have highlighted the effectiveness and ability of Machine Learning (ML) models in analyzing Remote Sensing (RS) data for this purpose. However, most of these models do not fully consider multi-source RS data for prediction, as processing these increases complexity and limits their accuracy and generalizability. In this study, we propose the Multi-Residual Attention-Based Multi-Stream 3D-ResNet-BiLSTM (MHRA-MS-3D-ResNet-BiLSTM) model, designed to integrate various RS data types, including Sentinel-1/2 imagery, Daymet climate data, and soil grid information, for improved county-level U.S. soybean yield prediction. Our model employs a multi-stream architecture to process diverse data types concurrently, capturing complex spatio-temporal features effectively. The 3D-ResNet component utilizes 3D convolutions and residual connections for pattern recognition, complemented by Bidirectional Long Short-Term Memory (BiLSTM) for enhanced long-term dependency learning by processing data arrangements in forward and backward directions. An attention mechanism further refines the model{\textquoteright}s focus by dynamically weighting the significance of different input features for efficient yield prediction. We trained the MHRA-MS-3D-ResNet-BiLSTM model using multi-source RS datasets from 2019 and 2020 and evaluated its performance with U.S. soybean yield data for 2021 and 2022. The results demonstrated the model{\textquoteright}s robustness and adaptability to unseen data, achieving an R2 of 0.82 and a Mean Absolute Percentage Error (MAPE) of 9% in 2021, and an R2 of 0.72 and MAPE of 12% in 2022. This performance surpassed some of the state-of-the-art models like 3D-ResNet-BiLSTM and MS-3D-ResNet-BiLSTM, and other traditional ML methods like Random Forest (RF), XGBoost, and LightGBM. These findings highlight the methodology{\textquoteright}s capability to handle multiple RS data types and its role in improving yield predictions, which can be helpful for sustainable agriculture.",
keywords = "3D-ResNet-biLSTM, Daymet, Google Earth Engine (GEE), Multi-Head Attention, Multi-Stream, Residual Attention, Sentinel-1, Sentinel-2, Soil Grid, soybean, yield prediction",
author = "Mahdiyeh Fathi and Reza Shah-Hosseini and Armin Moghimi and Hossein Arefi",
note = "Publisher Copyright: {\textcopyright} 2024 by the authors.",
year = "2024",
month = dec,
day = "31",
doi = "10.3390/rs17010107",
language = "English",
volume = "17",
journal = "Remote sensing",
issn = "2072-4292",
publisher = "Multidisciplinary Digital Publishing Institute",
number = "1",

}

Download

TY - JOUR

T1 - MHRA-MS-3D-ResNet-BiLSTM

T2 - A Multi-Head-Residual Attention-Based Multi-Stream Deep Learning Model for Soybean Yield Prediction in the U.S. Using Multi-Source Remote Sensing Data

AU - Fathi, Mahdiyeh

AU - Shah-Hosseini, Reza

AU - Moghimi, Armin

AU - Arefi, Hossein

N1 - Publisher Copyright: © 2024 by the authors.

PY - 2024/12/31

Y1 - 2024/12/31

N2 - Accurate prediction of soybean yield is important for safeguarding food security and improving agricultural management. Recent advances have highlighted the effectiveness and ability of Machine Learning (ML) models in analyzing Remote Sensing (RS) data for this purpose. However, most of these models do not fully consider multi-source RS data for prediction, as processing these increases complexity and limits their accuracy and generalizability. In this study, we propose the Multi-Residual Attention-Based Multi-Stream 3D-ResNet-BiLSTM (MHRA-MS-3D-ResNet-BiLSTM) model, designed to integrate various RS data types, including Sentinel-1/2 imagery, Daymet climate data, and soil grid information, for improved county-level U.S. soybean yield prediction. Our model employs a multi-stream architecture to process diverse data types concurrently, capturing complex spatio-temporal features effectively. The 3D-ResNet component utilizes 3D convolutions and residual connections for pattern recognition, complemented by Bidirectional Long Short-Term Memory (BiLSTM) for enhanced long-term dependency learning by processing data arrangements in forward and backward directions. An attention mechanism further refines the model’s focus by dynamically weighting the significance of different input features for efficient yield prediction. We trained the MHRA-MS-3D-ResNet-BiLSTM model using multi-source RS datasets from 2019 and 2020 and evaluated its performance with U.S. soybean yield data for 2021 and 2022. The results demonstrated the model’s robustness and adaptability to unseen data, achieving an R2 of 0.82 and a Mean Absolute Percentage Error (MAPE) of 9% in 2021, and an R2 of 0.72 and MAPE of 12% in 2022. This performance surpassed some of the state-of-the-art models like 3D-ResNet-BiLSTM and MS-3D-ResNet-BiLSTM, and other traditional ML methods like Random Forest (RF), XGBoost, and LightGBM. These findings highlight the methodology’s capability to handle multiple RS data types and its role in improving yield predictions, which can be helpful for sustainable agriculture.

AB - Accurate prediction of soybean yield is important for safeguarding food security and improving agricultural management. Recent advances have highlighted the effectiveness and ability of Machine Learning (ML) models in analyzing Remote Sensing (RS) data for this purpose. However, most of these models do not fully consider multi-source RS data for prediction, as processing these increases complexity and limits their accuracy and generalizability. In this study, we propose the Multi-Residual Attention-Based Multi-Stream 3D-ResNet-BiLSTM (MHRA-MS-3D-ResNet-BiLSTM) model, designed to integrate various RS data types, including Sentinel-1/2 imagery, Daymet climate data, and soil grid information, for improved county-level U.S. soybean yield prediction. Our model employs a multi-stream architecture to process diverse data types concurrently, capturing complex spatio-temporal features effectively. The 3D-ResNet component utilizes 3D convolutions and residual connections for pattern recognition, complemented by Bidirectional Long Short-Term Memory (BiLSTM) for enhanced long-term dependency learning by processing data arrangements in forward and backward directions. An attention mechanism further refines the model’s focus by dynamically weighting the significance of different input features for efficient yield prediction. We trained the MHRA-MS-3D-ResNet-BiLSTM model using multi-source RS datasets from 2019 and 2020 and evaluated its performance with U.S. soybean yield data for 2021 and 2022. The results demonstrated the model’s robustness and adaptability to unseen data, achieving an R2 of 0.82 and a Mean Absolute Percentage Error (MAPE) of 9% in 2021, and an R2 of 0.72 and MAPE of 12% in 2022. This performance surpassed some of the state-of-the-art models like 3D-ResNet-BiLSTM and MS-3D-ResNet-BiLSTM, and other traditional ML methods like Random Forest (RF), XGBoost, and LightGBM. These findings highlight the methodology’s capability to handle multiple RS data types and its role in improving yield predictions, which can be helpful for sustainable agriculture.

KW - 3D-ResNet-biLSTM

KW - Daymet

KW - Google Earth Engine (GEE)

KW - Multi-Head Attention

KW - Multi-Stream

KW - Residual Attention

KW - Sentinel-1

KW - Sentinel-2

KW - Soil Grid

KW - soybean

KW - yield prediction

UR - http://www.scopus.com/inward/record.url?scp=85214521866&partnerID=8YFLogxK

U2 - 10.3390/rs17010107

DO - 10.3390/rs17010107

M3 - Article

AN - SCOPUS:85214521866

VL - 17

JO - Remote sensing

JF - Remote sensing

SN - 2072-4292

IS - 1

M1 - 107

ER -

Von denselben Autoren