Soybean Yield Estimation Using Improved Deep Learning Models With Integrated Multisource and Multitemporal Remote Sensing Data

Accurate soybean yield estimation is critically imperative for modern agricultural systems amid escalating global food security pressures, yet conventional methodologies are constrained for large-scale high-frequency monitoring. To address this, an innovative deep learning framework, TransBiHGRU-PSO...

Full description

Saved in:
Bibliographic Details
Main Authors: Jian Li, Junrui Kang, Ji Qi, Jian Lu, Hongkun Fu, Baoqi Liu, Xinglei Lin, Jiawei Zhao, Hengxu Guan, Jing Chang, Zhihan Liu
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11079987/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849340822645899264
author Jian Li
Junrui Kang
Ji Qi
Jian Lu
Hongkun Fu
Baoqi Liu
Xinglei Lin
Jiawei Zhao
Hengxu Guan
Jing Chang
Zhihan Liu
author_facet Jian Li
Junrui Kang
Ji Qi
Jian Lu
Hongkun Fu
Baoqi Liu
Xinglei Lin
Jiawei Zhao
Hengxu Guan
Jing Chang
Zhihan Liu
author_sort Jian Li
collection DOAJ
description Accurate soybean yield estimation is critically imperative for modern agricultural systems amid escalating global food security pressures, yet conventional methodologies are constrained for large-scale high-frequency monitoring. To address this, an innovative deep learning framework, TransBiHGRU-PSO, is proposed for precise large-scale soybean yield estimation via effective fusion of multisource multitemporal remote sensing data, emphasizing robust and accurate estimation even with anomalous yield data. This framework synergistically integrates an optimized bidirectional hierarchical gated recurrent unit (BiHGRU), a Transformer encoder, and a novel Greenness and Water Content Composite Index, with critical parameters optimized by particle swarm optimization (PSO). County-level yield data from 12 U.S. states were used, supplemented by multitemporal remote sensing datasets (MODIS surface reflectance, vegetation indices, and environmental variables). Empirical analyses showed that TransBiHGRU-PSO demonstrated improved estimation capability and generalizability compared to multiple benchmark models. Notably, with anomalous yield data retained, the model achieved solid test set performance [coefficient of determination (<italic>R</italic><sup>2</sup>) of 0.71 and root-mean-square error (RMSE) of 4.2812 bushels/acre]. Compared to the best traditional machine learning model (support vector regression), <italic>R</italic><sup>2</sup> increased by 52.96% and RMSE decreased by 26.05%, and relative to the best deep learning baseline model (long short-term memory), <italic>R</italic><sup>2</sup> and RMSE improved by 7.04% and 7.04%, respectively. Furthermore, validation of interannual stability (2008&#x2013;2018, anomalies retained) revealed a mean <italic>R</italic><sup>2</sup> of 0.70 and a mean RMSE of 4.4701 bushels/acre, affirming its consistency under complex real-world conditions. This TransBiHGRU-PSO algorithmic framework, combined with multisource and multitemporal remote sensing data, offers a valuable exploration for large-scale soybean yield estimation.
format Article
id doaj-art-e7dabdfd5563435d897b1cf182b1d14a
institution Kabale University
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-e7dabdfd5563435d897b1cf182b1d14a2025-08-20T03:43:47ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118184501847710.1109/JSTARS.2025.358891711079987Soybean Yield Estimation Using Improved Deep Learning Models With Integrated Multisource and Multitemporal Remote Sensing DataJian Li0https://orcid.org/0009-0008-7749-0286Junrui Kang1https://orcid.org/0009-0008-8603-0492Ji Qi2Jian Lu3Hongkun Fu4Baoqi Liu5Xinglei Lin6Jiawei Zhao7Hengxu Guan8Jing Chang9https://orcid.org/0000-0003-4825-6382Zhihan Liu10College of Information Technology, Jilin Agricultural University, Changchun, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun, ChinaCollege of Engineering and Technology, Jilin Agricultural University, Changchun, ChinaCollege of Agronomy, Jilin Agricultural University, Changchun, ChinaCollege of Agronomy, Jilin Agricultural University, Changchun, ChinaNortheast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun, ChinaAccurate soybean yield estimation is critically imperative for modern agricultural systems amid escalating global food security pressures, yet conventional methodologies are constrained for large-scale high-frequency monitoring. To address this, an innovative deep learning framework, TransBiHGRU-PSO, is proposed for precise large-scale soybean yield estimation via effective fusion of multisource multitemporal remote sensing data, emphasizing robust and accurate estimation even with anomalous yield data. This framework synergistically integrates an optimized bidirectional hierarchical gated recurrent unit (BiHGRU), a Transformer encoder, and a novel Greenness and Water Content Composite Index, with critical parameters optimized by particle swarm optimization (PSO). County-level yield data from 12 U.S. states were used, supplemented by multitemporal remote sensing datasets (MODIS surface reflectance, vegetation indices, and environmental variables). Empirical analyses showed that TransBiHGRU-PSO demonstrated improved estimation capability and generalizability compared to multiple benchmark models. Notably, with anomalous yield data retained, the model achieved solid test set performance [coefficient of determination (<italic>R</italic><sup>2</sup>) of 0.71 and root-mean-square error (RMSE) of 4.2812 bushels/acre]. Compared to the best traditional machine learning model (support vector regression), <italic>R</italic><sup>2</sup> increased by 52.96% and RMSE decreased by 26.05%, and relative to the best deep learning baseline model (long short-term memory), <italic>R</italic><sup>2</sup> and RMSE improved by 7.04% and 7.04%, respectively. Furthermore, validation of interannual stability (2008&#x2013;2018, anomalies retained) revealed a mean <italic>R</italic><sup>2</sup> of 0.70 and a mean RMSE of 4.4701 bushels/acre, affirming its consistency under complex real-world conditions. This TransBiHGRU-PSO algorithmic framework, combined with multisource and multitemporal remote sensing data, offers a valuable exploration for large-scale soybean yield estimation.https://ieeexplore.ieee.org/document/11079987/Deep learning (DL)greenness and water content composite index (GWCCI)multisource remote sensing datayield estimation
spellingShingle Jian Li
Junrui Kang
Ji Qi
Jian Lu
Hongkun Fu
Baoqi Liu
Xinglei Lin
Jiawei Zhao
Hengxu Guan
Jing Chang
Zhihan Liu
Soybean Yield Estimation Using Improved Deep Learning Models With Integrated Multisource and Multitemporal Remote Sensing Data
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Deep learning (DL)
greenness and water content composite index (GWCCI)
multisource remote sensing data
yield estimation
title Soybean Yield Estimation Using Improved Deep Learning Models With Integrated Multisource and Multitemporal Remote Sensing Data
title_full Soybean Yield Estimation Using Improved Deep Learning Models With Integrated Multisource and Multitemporal Remote Sensing Data
title_fullStr Soybean Yield Estimation Using Improved Deep Learning Models With Integrated Multisource and Multitemporal Remote Sensing Data
title_full_unstemmed Soybean Yield Estimation Using Improved Deep Learning Models With Integrated Multisource and Multitemporal Remote Sensing Data
title_short Soybean Yield Estimation Using Improved Deep Learning Models With Integrated Multisource and Multitemporal Remote Sensing Data
title_sort soybean yield estimation using improved deep learning models with integrated multisource and multitemporal remote sensing data
topic Deep learning (DL)
greenness and water content composite index (GWCCI)
multisource remote sensing data
yield estimation
url https://ieeexplore.ieee.org/document/11079987/
work_keys_str_mv AT jianli soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata
AT junruikang soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata
AT jiqi soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata
AT jianlu soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata
AT hongkunfu soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata
AT baoqiliu soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata
AT xingleilin soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata
AT jiaweizhao soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata
AT hengxuguan soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata
AT jingchang soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata
AT zhihanliu soybeanyieldestimationusingimproveddeeplearningmodelswithintegratedmultisourceandmultitemporalremotesensingdata