Deep Transfer Learning for UAV-Based Cross-Crop Yield Prediction in Root Crops

Authors

• Suraj A. Yadav, Mississippi State University
• Yanbo Huang, USDA Agricultural Research Service
• Kenny Q. Zhu, University of Texas
• Rayyan Haque, University of Texas
• Wyatt Young, USDA Agricultural Research Service
• Lorin Harvey, Mississippi State University
• Mark Hall, Mississippi State University
• Xin Zhang, University of Georgia
• Nuwan K. Wijewardane, Mississippi State University
• Ruijun Qin, Oregon State University
• Max Feldman, USDA Agricultural Research Service
• Haibo Yao, USDA Agricultural Research Service
• John P. Brooks, USDA Agricultural Research Service

ORCID

Yadav: https://orcid.org/0000-0002-8158-7207; Huang: https://orcid.org/0000-0002-1409-8868; Young: https://orcid.org/0009-0005-9736-1228; Harvey: https://orcid.org/0000-0002-5571-7792; Wijewardane: https://orcid.org/0000-0001-8962-9451; Qin: https://orcid.org/0000-0003-3998-3385; Feldman: https://orcid.org/0000-0002-5415-4326; Brooks: https://orcid.org/0000-0002-6142-6430

MSU Affiliation

College of Agriculture and Life Sciences; Department of Plant and Soil Sciences; Department of Agricultural and Biological Engineering; James Worth Bagley College of Engineering

Creation Date

2026-06-30

Abstract

Limited annotated data often constrain accurate yield prediction in underrepresented crops. To address this challenge, we developed a cross-crop deep transfer learning (TL) framework that leverages potato (Solanum tuberosum L.) as the source domain to predict sweet potato (Ipomoea batatas L.) yield using multi-temporal uncrewed aerial vehicle (UAV)-based multispectral imagery. A hybrid convolutional–recurrent neural network (CNN–RNN–Attention) architecture was implemented with a robust parameter-based transfer strategy to ensure temporal alignment and feature-space consistency across crops. Cross-crop feature migration analysis showed that predictors capturing canopy vigor, structure, and soil–vegetation contrast exhibited the highest distributional similarity between potato and sweet potato. In comparison, pigment-sensitive and agronomic predictors were less transferable. These robustness patterns were reflected in model performance, as all architectures showed substantial improvement when moving from the minimal 3 predictor subset to the 5–7 predictor subsets, where the most transferable indices were introduced. The hybrid CNN–RNN–Attention model achieved peak accuracy (𝑅2≈0.64 and RMSE ≈ 18%) using time-series data up to the tuberization stage with only 7 predictors. In contrast, convolutional neural network (CNN), bidirectional gated recurrent unit (BiGRU), and bidirectional long short-term memory (BiLSTM) baseline models required 11–13 predictors to achieve comparable performance and often showed reduced or unstable accuracy at higher dimensionality due to redundancy and domain-shift amplification. Two-way ANOVA further revealed that cover crop type significantly influenced yield, whereas nitrogen rate and the interaction term were not significant. Overall, this study demonstrates that combining robustness-aware feature design with hybrid deep TL model enables accurate, data-efficient, and physiologically interpretable yield prediction in sweet potato, offering a scalable pathway for applying TL in other underrepresented root and tuber crops.

Publication Date

12-17-2025

Publication Title

Remote Sensing

Publisher

MDPI

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

Rights

© 2025 The Authors

Recommended Citation

Yadav, S. A., Huang, Y., Zhu, K. Q., Haque, R., Young, W., Harvey, L., Hall, M., Zhang, X., Wijewardane, N. K., Qin, R., Feldman, M., Yao, H., & Brooks, J. P. (2025). Deep Transfer Learning for UAV-Based Cross-Crop Yield Prediction in Root Crops. Remote Sensing, 17(24), 4054. https://doi.org/10.3390/rs17244054