Enhancing corn yield prediction: Optimizing data quality or model complexity?
Field-scale corn yield prediction before harvest can assist farmers in better organizing their resources. Machine learning-based pipelines for analyzing remote sensing imagery offer an efficient solution to this problem. However, the cost of data acquisition and training requirements for machine or...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2024-12-01
|
| Series: | Smart Agricultural Technology |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772375524002764 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850106939073101824 |
|---|---|
| author | Yuting Zhou Shengfang Ma Huihui Zhang Sathyanarayanan Aakur |
| author_facet | Yuting Zhou Shengfang Ma Huihui Zhang Sathyanarayanan Aakur |
| author_sort | Yuting Zhou |
| collection | DOAJ |
| description | Field-scale corn yield prediction before harvest can assist farmers in better organizing their resources. Machine learning-based pipelines for analyzing remote sensing imagery offer an efficient solution to this problem. However, the cost of data acquisition and training requirements for machine or deep learning models depend on various factors, such as equipment (multispectral vs. RGB sensors) and the ability to predict yield from observations across growth stages. In this study, we aim to provide a comprehensive analysis of the effectiveness of traditional ensemble learning methods (Random Forest and Gradient Boosting) and deep learning models (ResNet 18, ResNet34, and ViT) in predicting corn yield across deficit and fully irrigated fields using UAV-based RGB and multispectral imagery. The performance of these models was examined across early, middle, and late growth stages, considering both computational complexity and accuracy. We also developed a novel shallow CNN framework called SimRes, inspired by the ResNet framework but tailored for streamlined efficiency and simplicity for yield prediction. Extensive quantitative analysis demonstrated that the customized SimRes performed as well as deep learning baselines but with faster computing times, while traditional approaches, such as Random Forests and Gradient Boosting exhibited marginally smaller R-squared values. Models utilizing multispectral data outperformed models using RGB, albeit with variations across growth stages. Deep learning methods performed better than ensemble learning methods in the early and late growth stages using RGB, while performance became comparable in the middle stage. These results underscore the importance of additional information or more complex models to enhance prediction accuracy alongside a trade-off between computational complexity and accuracy. This research provides valuable insights for optimizing corn yield prediction across different growth stages, informing agricultural management and harvest planning decisions. |
| format | Article |
| id | doaj-art-214224ad4b4d42dd872929a76f53ec0d |
| institution | OA Journals |
| issn | 2772-3755 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Smart Agricultural Technology |
| spelling | doaj-art-214224ad4b4d42dd872929a76f53ec0d2025-08-20T02:38:42ZengElsevierSmart Agricultural Technology2772-37552024-12-01910067110.1016/j.atech.2024.100671Enhancing corn yield prediction: Optimizing data quality or model complexity?Yuting Zhou0Shengfang Ma1Huihui Zhang2Sathyanarayanan Aakur3Department of Geography, Oklahoma State University, Stillwater, OK, USA; Corresponding authors.Oklahoma Water Resources Center, Oklahoma State University, Stillwater, OK, USWater Management and Systems Research Unit, United States Department of Agriculture, Agricultural Research Service, Fort Collins, CO, US; Corresponding authors.Department of Computer Science and Software Engineering, Auburn University, Auburn, AL, USAField-scale corn yield prediction before harvest can assist farmers in better organizing their resources. Machine learning-based pipelines for analyzing remote sensing imagery offer an efficient solution to this problem. However, the cost of data acquisition and training requirements for machine or deep learning models depend on various factors, such as equipment (multispectral vs. RGB sensors) and the ability to predict yield from observations across growth stages. In this study, we aim to provide a comprehensive analysis of the effectiveness of traditional ensemble learning methods (Random Forest and Gradient Boosting) and deep learning models (ResNet 18, ResNet34, and ViT) in predicting corn yield across deficit and fully irrigated fields using UAV-based RGB and multispectral imagery. The performance of these models was examined across early, middle, and late growth stages, considering both computational complexity and accuracy. We also developed a novel shallow CNN framework called SimRes, inspired by the ResNet framework but tailored for streamlined efficiency and simplicity for yield prediction. Extensive quantitative analysis demonstrated that the customized SimRes performed as well as deep learning baselines but with faster computing times, while traditional approaches, such as Random Forests and Gradient Boosting exhibited marginally smaller R-squared values. Models utilizing multispectral data outperformed models using RGB, albeit with variations across growth stages. Deep learning methods performed better than ensemble learning methods in the early and late growth stages using RGB, while performance became comparable in the middle stage. These results underscore the importance of additional information or more complex models to enhance prediction accuracy alongside a trade-off between computational complexity and accuracy. This research provides valuable insights for optimizing corn yield prediction across different growth stages, informing agricultural management and harvest planning decisions.http://www.sciencedirect.com/science/article/pii/S2772375524002764UAVMultispectralRGBYield predictionDeep learningVision transformer |
| spellingShingle | Yuting Zhou Shengfang Ma Huihui Zhang Sathyanarayanan Aakur Enhancing corn yield prediction: Optimizing data quality or model complexity? Smart Agricultural Technology UAV Multispectral RGB Yield prediction Deep learning Vision transformer |
| title | Enhancing corn yield prediction: Optimizing data quality or model complexity? |
| title_full | Enhancing corn yield prediction: Optimizing data quality or model complexity? |
| title_fullStr | Enhancing corn yield prediction: Optimizing data quality or model complexity? |
| title_full_unstemmed | Enhancing corn yield prediction: Optimizing data quality or model complexity? |
| title_short | Enhancing corn yield prediction: Optimizing data quality or model complexity? |
| title_sort | enhancing corn yield prediction optimizing data quality or model complexity |
| topic | UAV Multispectral RGB Yield prediction Deep learning Vision transformer |
| url | http://www.sciencedirect.com/science/article/pii/S2772375524002764 |
| work_keys_str_mv | AT yutingzhou enhancingcornyieldpredictionoptimizingdataqualityormodelcomplexity AT shengfangma enhancingcornyieldpredictionoptimizingdataqualityormodelcomplexity AT huihuizhang enhancingcornyieldpredictionoptimizingdataqualityormodelcomplexity AT sathyanarayananaakur enhancingcornyieldpredictionoptimizingdataqualityormodelcomplexity |