Sweet Potato Yield Prediction Using Machine Learning Based on Multispectral Images Acquired from a Small Unmanned Aerial Vehicle

Kriti Singh, Yanbo Huang (), Wyatt Young, Lorin Harvey, Mark Hall, Xin Zhang, Edgar Lobaton, Johnie Jenkins and Mark Shankle
Additional contact information
Kriti Singh: Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, NC 27606, USA
Yanbo Huang: USDA-ARS Genetics and Sustainable Agriculture Research Unit, Mississippi State, MS 39762, USA
Wyatt Young: USDA-ARS Genetics and Sustainable Agriculture Research Unit, Mississippi State, MS 39762, USA
Lorin Harvey: Pontotoc Ridge-Flatwoods Branch Experiment Station, Mississippi State University, Pontotoc, MS 38863, USA
Mark Hall: Pontotoc Ridge-Flatwoods Branch Experiment Station, Mississippi State University, Pontotoc, MS 38863, USA
Xin Zhang: Department of Agricultural and Biological Engineering, Mississippi State University, Mississippi State, MS 39762, USA
Edgar Lobaton: Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, NC 27606, USA
Johnie Jenkins: USDA-ARS Genetics and Sustainable Agriculture Research Unit, Mississippi State, MS 39762, USA
Mark Shankle: Pontotoc Ridge-Flatwoods Branch Experiment Station, Mississippi State University, Pontotoc, MS 38863, USA

Agriculture, 2025, vol. 15, issue 4, 1-23

Abstract: Accurate prediction of sweet potato yield is crucial for effective crop management. This study investigates the use of vegetation indices (VIs) extracted from multispectral images acquired by a small unmanned aerial vehicle (UAV) throughout the growing season, along with in situ-measured plant physiological parameters, to predict sweet potato yield. The data acquisition process through UAV field imaging is discussed in detail along with the extraction process for the multispectral bands that we use as features. The experiment is designed with a combination of different nitrogen application rates and cover crop treatments. The dependence of VIs and crop physiological parameters, such as leaf chlorophyll content, plant biomass, vine length, and leaf nitrogen content, on yield is evaluated through feature selection methods and model performance. Classical machine learning (ML) approaches and tree-based algorithms, like XGBoost and Random Forest, are implemented. Additionally, a soft-voting ML model ensemble approach is employed to improve performance of yield prediction. Individual models are trained and tested for different cover crop and nitrogen treatments to capture the relationships between the treatments and the target yield variable. The performance of the ML algorithms is evaluated using various popular model performance metrics like R 2 , RMSE, and MAE. Through modelling the data for cover crops and nitrogen treatment rates using individual models, the relationships and effects of different treatments on yield are explored. Important VIs useful for the study are identified through feature selection and model performance evaluation.

Keywords: sweet potato; yield; remote sensing; unmanned aerial vehicle; machine learning (search for similar items in EconPapers)
JEL-codes: Q1 Q10 Q11 Q12 Q13 Q14 Q15 Q16 Q17 Q18 (search for similar items in EconPapers)
Date: 2025
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/2077-0472/15/4/420/pdf (application/pdf)
https://www.mdpi.com/2077-0472/15/4/420/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jagris:v:15:y:2025:i:4:p:420-:d:1592804

Access Statistics for this article

Agriculture is currently edited by Ms. Leda Xuan

More articles in Agriculture from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().