A Three-Dimensional Conceptual Model for Estimating the Above-Ground Biomass of Winter Wheat Using Digital and Multispectral Unmanned Aerial Vehicle Images at Various Growth Stages

Abstract

The timely and accurate estimation of above-ground biomass (AGB) is crucial for indicating crop growth status, assisting management decisions, and predicting grain yield. Unmanned aerial vehicle (UAV) remote sensing technology is a promising approach for monitoring crop biomass. However, the determination of winter wheat AGB based on canopy reflectance is affected by spectral saturation effects. Thus, constructing a generic model for accurately estimating winter wheat AGB using UAV data is significant. In this study, a three-dimensional conceptual model (3DCM) for estimating winter wheat AGB was constructed using plant height (PH) and fractional vegetation cover (FVC). Compared with both the traditional vegetation index model and the traditional multi-feature combination model, the 3DCM yielded the best accuracy for the jointing stage (based on RGB data: coefficient of determination (R2) = 0.82, normalized root mean square error (nRMSE) = 0.2; based on multispectral (MS) data: R2 = 0.84, nRMSE = 0.16), but the accuracy decreased significantly when the spike organ appeared. Therefore, the spike number (SN) was added to create a new three-dimensional conceptual model (n3DCM). Under different growth stages and UAV platforms, the n3DCM (RGB: R2 = 0.73–0.85, nRMSE = 0.17–0.23; MS: R2 = 0.77–0.84, nRMSE = 0.17–0.23) remarkably outperformed the traditional multi-feature combination model (RGB: R2 = 0.67–0.88, nRMSE = 0.15–0.25; MS: R2 = 0.60–0.77, nRMSE = 0.19–0.26) for the estimation accuracy of the AGB. This study suggests that the n3DCM has great potential in resolving spectral errors and monitoring growth parameters, which could be extended to other crops and regions for AGB estimation and field-based high-throughput phenotyping.

Optimizing window size and directional parameters of GLCM texture features for estimating rice AGB based on UAVs multispectral imagery

“…The result showed that the different performance between SFs and TFs was stage-specific. During the early stages of crop growth, the high leaf-stem density plays a significant role in determining a substantial portion of the AGB ( Zhu et al., 2023 ). The SFs and TFs can reflect the reflectance attribute and spatial variation of rice canopy well, and the correlation between them is high ( Bai et al., 2021 ).…”

“…When adjusting ntree to a sufficiently large value, it primarily impacts the modeling time rather than the modeling accuracy ( Wang et al., 2016 ; Zhang et al., 2021 ). Therefore, following its application in other studies, we set ntree to 1,000 ( Li et al., 2019 ; Zhu et al., 2023 ). On the other hand, the value of mtry significantly affects the modeling accuracy of RF.…”

“…Extensive research has been conducted utilizing spectral information from UAVs imagery to monitor crop AGB, and it has achieved acceptable monitoring accuracy ( Han et al., 2019 ; Maimaitijiang et al., 2019 ; Wang F. et al., 2022 ; Wang et al., 2023 ). However, spectral information is susceptible to interference from water-soil (and weed) background noise during the early stages, while the dense canopy coverage at the later stages can result in spectral saturation ( Li et al., 2022 ; Wang F. et al., 2022 ; Wang Q. et al., 2022 ; Wang W. et al., 2022 ; Zhu et al., 2023 ). These factors impose constraints on further advancements in monitoring accuracy.…”

Field-Scale Winter Wheat Growth Prediction Applying Machine Learning Methods with Unmanned Aerial Vehicle Imagery and Soil Properties

“…Furthermore, wheat cultivated within infertile acidic soil experiences a reduced protein content and growth rate and lower yields, which result in reduced profits. Soil elements such as phosphorus (P), potassium (K), magnesium (Mg), calcium (Ca), sodium (Na), nitrogen (N), and pH are vital for crop growth and often exist in low concentrations in arid and semi-arid environments [15][16][17][18]. Deficiencies of N, P, and K in soil affect wheat growth and yield drastically [19].…”

“…The common modelling approaches include parametric and non-parametric regression for crop biophysical parameter estimation. These include partial least squares regression (PLSR), random forest (RF), support vector machine (SVM), extreme gradient boosting (Xgboost), conditional inference forest (CI-forest), artificial neural network (ANN), least squares linear regression (LSLR), multiple linear regression (LR), neural network (NN), decision tree (DT), regression tree (RegT), K-nearest neighbour (KNN), boost tree (BST), and bagging tree (BagT) ensemble learning algorithms [7,18,30,[34][35][36]. PLSR provides a high level of interpretability and can overcome problems of collinearity in modelling, enhancing the accuracy of the model [9,17].…”