Application of UAV multispectral imaging to monitor soybean growth with yield prediction through machine learning

Authors: SHAMMI, SADIA - Oak Ridge Institute For Science And Education (ORISE); Huang, Yanbo; Feng, Gary; Tewolde, Haile; ZHANG, XIN - Mississippi State University; Jenkins, Johnie; SHANKLE, MARK - Mississippi State University

Submitted to: Agronomy Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/25/2024
Publication Date: 3/26/2024
Citation: Shammi, S.A., Huang, Y., Feng, G.G., Tewolde, H., Zhang, X., Jenkins, J.N., Shankle, M. 2024. Application of UAV multispectral imaging to monitor soybean growth with yield prediction through machine learning. Agronomy Journal. 14(672):1-18. https://doi.org/10.3390/agronomy14040672.
DOI: https://doi.org/10.3390/agronomy14040672

Interpretive Summary: Remote sensing has been developed for crop production management, especially UAV remote sensing in recent years. In this study the scientists in USDA ARS and Mississippi State university collaboratively developed UAV multispectral remote sensing along soybean growth in field with different fertilizer treatments and cover crops. The results indicate that the vegetation indices calculated for the extracted spectral data from UAV acquired multispectral images can acceptably predict soybean yield using machine learning algorithms at different growth stages with the interactions of fertilizer treatments and cover crops. The development of this study provides a scheme of UAV remote sensing using machine learning for multi-stage crop yield prediction along the crop growth process to provide useful information for decision support in crop production management.

Technical Abstract: Application of remote sensing, which is non-destructive and cost-efficient, has been widely used in crop monitoring and management. This study used a built-in multispectral imager on a small unmanned aerial vehicle (UAV) to capture multispectral images in five different spectral bands (blue, green, red, red edge, and near-infrared), instead of satellite captured data, to monitor soybean growth in the field. The field experiment was conducted in a soybean field of Mississippi State University Experiment Station near Pontotoc, MS, USA. The experiment consisted of five cover crops (Cereal Rye, Vetch, Wheat, Mustard plus Cereal Rye, and native vegetation) planted in the winter and three fertilizer treatments (Fertilizer, Poultry Liter, and None) applied before planting the soybean. During the soybean growing season in 2022, eight UAV imaging flyovers were conducted spread across the growth season. UAV image-derived vegetation indices (VIs) coupled with machine learning (ML) models were computed for characterizing soybean growth at different stages across the season. The aim of this study focuses on monitoring soybean growth to predict yield using 14 VIs including CC (Canopy cover), NDVI (Normalized difference vegetation index), GNDVI (Green normalized difference vegetation index), EVI2 (Enhanced Vegetation Index 2), and others. Different machine learning algorithms including Linear Regression (LR), Support Vector Machine (SVM), and Random Forest (RF) are used for this purpose. The stage of initial pod development) was shown as the best predictability for earliest soybean yield prediction. CC, NDVI, and NAVI were shown as the best VIs for yield prediction. The RMSE is found to be about 134.5 to 511.11 kg ha-1 in different yield models, whereas 605.26 to 685.96 kg ha-1 in cross validated models. Due to the limited number of training and testing samples in K-fold cross validation, the models’ results changed to some extent. Nevertheless, the result of this study will be useful for the application of UAV remote sensing to provide information for soybean production and management. This study demonstrates that VIs coupled with ML models can be used in multistage soybean yield prediction at farm scale, even with limited number of training samples.