AUTHOR=Fei Shuaipeng , Hassan Muhammad Adeel , Ma Yuntao , Shu Meiyan , Cheng Qian , Li Zongpeng , Chen Zhen , Xiao Yonggui 

TITLE=Entropy Weight Ensemble Framework for Yield Prediction of Winter Wheat Under Different Water Stress Treatments Using Unmanned Aerial Vehicle-Based Multispectral and Thermal Data

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 12 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2021.730181

DOI=10.3389/fpls.2021.730181

ISSN=1664-462X

ABSTRACT=Crop breeding programs generally perform early field assessments of candidates selection based on the primary traits such as grain yield (GY). Traditional methods of yield assessment are costly, inefficient, and considered as bottleneck in modern precision agriculture. Recent advances in unmanned aerial vehicle (UAV) and sensors development have opened new avenue for data acquisition in cost-effective and rapid manner. We evaluated UAV-based multispectral and thermal images for in-season GY prediction using 30 winter wheat genotypes under three water treatments. For this, multispectral vegetation indices (VIs) and temperature index were calculated and selected by gray correlation analysis (GRA) at each growth stage i.e. jointing, booting, heading, flowering, grain filling and maturity to reduce the data dimension. The elastic net (ENET) regression was developed by using selected features as input variables for yield prediction. Whereas entropy weight fusion (EWF) method was used to combine the predicted GY values from multiple growth stages. Results showed that grain filling stage was the optimal stage for predicting of GY with R2=0.667, RMSE=0.881 t ha-1, RRMSE=15.211% and MAE=0.721 t ha-1. The EWF model outperformed at all the individual growth stages. The best prediction result (R2=0.729, RMSE=0.831 t ha-1, RRMSE=14.347% and MAE=0.684 t ha-1) was achieved through combining the predicted values of first five growth stages. This study suggested that the fusion of UAV-based multispectral and thermal infrared data within an ENET-EWF framework can provide precise and robust prediction of wheat yield.