Estimation of grain yield in wheat using source–sink datasets derived from RGB and thermal infrared imaging

Abstract

Timely and efficient monitoring of crop aboveground biomass (AGB) and grain yield (GY) forecasting before harvesting are critical for improving crop yields and ensuring food security in precision agriculture. The purpose of this study is to explore the potential of fusing source–sink‐level color, texture, and temperature values extracted from RGB images and thermal images based on proximal sensing technology to improve grain yield prediction. High‐quality images of wheat from flowering to maturity under different treatments of nitrogen application were collected using proximal sensing technology over a 2‐year trial. Numerous variables based on source and sink organs were extracted from the acquired subsample images, including 30 color features, 10 texture features, and two temperature values. The principal component analysis (PCA), least absolute shrinkage and selection operator (LASSO), and recursive feature elimination (RFE) were used to screen variables. Support vector regression (SVR) and random forest (RF) were applied to establish AGB estimation models, and the GY prediction models were built by RF. The source dataset and sink dataset performed differently on AGB and GY estimation, but the combined source–sink dataset performed best for estimating both AGB and GY. Based on the source–sink dataset, the LASSO‐RF model was the best combination for predicting AGB and GY, with the coefficient of determination (R2) of 0.85 and 0.86, root mean square error (RMSE) of 1179.09 and 609.61 kg ha−1, and the ratio of performance to deviation (RPD) of 2.10 and 2.45, respectively. This study demonstrates that the multivariate eigenvalues of both source and sink organs have the potential to predict wheat yield and that the combination of machine learning models and variable selection methods can significantly affect the accuracy of yield prediction models and achieve effective monitoring of crop growth at late reproductive stages.

Advancements in Utilizing Image-Analysis Technology for Crop-Yield Estimation

“…The highest prediction result was obtained using an integrated model based on SVM, GP, LRR, and RF, with an R 2 of 0.78, which was superior to a single machine learning algorithm and independent variables without feature optimization. In terms of data-fusion yield calculation, the main approach is to use multi-sensor and multichannel data to establish a wheat-yield calculation model, and many more results have been achieved than with a single dimension, such as those of Fei et al [38], Li et al [62], Sharif et al [63], and Ma et al [64], who have studied some wheat-yield calculation models based on the fusion of multi-channel data such as RGB images, multispectral, thermal infrared images, and meteorological data. The results obtained through multiple machine learning algorithms are superior to single-channel modeling, and the calculation accuracy and robustness were more advantageous.…”

Estimation of Agronomic Characters of Wheat Based on Variable Selection and Machine Learning Algorithms

“…For example, Wang's results show that the RF variable-screening algorithm exhibits good performance in processing wheat SPAD data [30]. The experimental results of Li et al indicated that the estimation accuracy in the wheat yield prediction model could be improved by using the LASSO variable-selection algorithm [75].…”