Estimating Leaf Nitrogen Content in Wheat Using Multimodal Features Extracted from Canopy Spectra

Abstract

The leaf nitrogen content (LNC) of wheat is one of key bases for wheat nitrogen fertilizer management and nutritional diagnosis, which is of great significance to the sustainable development of precision agriculture. The canopy spectrum provides an effective way to monitor the nitrogen content of wheat. Previous studies have shown that features extracted from the canopy spectrum, such as vegetation indices (VIs) and band positions (BPs), have successfully achieved the monitoring of crop nitrogen nutrition. However, the features mentioned above are spectral features extracted on the basis of linear or nonlinear combination models with a simple structure, which limits the general applicability of the model. In addition, models based on spectral features are prone to overfitting, which also reduces the accuracy of the model. Therefore, we propose an estimation model based on multimodal features (convolutional features and VIs, BPs) of the canopy spectrum, which aim to improve accuracy in estimating wheat LNC. Among these, the convolutional features (CFs) extracted by the designed convolutional neural network represent the deep semantic information of the canopy reflection spectrum, which can make up for the lack of robustness of the spectral features. The results showed that the accuracy of the model based on the fusion features (VIs + BPs + CFs) was higher than that of the feature of single modality. Moreover, the particle swarm optimization–support vector regression (PSO-SVR) model based on multimodal features had the best prediction effect (R2 = 0.896, RMSE = 0.188 for calibration, R2 = 0.793, RMSE = 0.408 for validation). Therefore, the method proposed in this study could improve performance in the estimation of wheat LNC, which provides technical support for wheat nitrogen nutrition monitoring.

The prediction model of nitrogen nutrition in cotton canopy leaves based on hyperspectral visible‐near infrared band feature fusion

“…(4) Improve model stability: Fused features can reduce the dependence of the model on a single feature, improve the stability and robustness of the model, and reduce the risk of model over‐fitting. (5) Improve the interpretability of the model: fused features can combine different types of feature information, improve the interpretability of the model, and better understand the prediction process and results of the model [ 33–35 ] …”

“…[32] The advantages of building prediction model by fusing features are Improve the interpretability of the model: fused features can combine different types of feature information, improve the interpretability of the model, and better understand the prediction process and results of the model. [33][34][35] The information fusion type used in this paper was feature level…”

Visible and near-infrared spectroscopy and deep learning application for the qualitative and quantitative investigation of nitrogen status in cotton leaves

“…When using WPLS for wavelength selection, a PLS regression model is first established, and each variable’s regression coefficient was calculated. The wavelengths with the larger absolute value of the regression coefficient at the crest and trough were selected ( Mehmood et al., 2012 ). Saliency map is a popular method for computing the contribution of each variable to the model performance.…”

“…Relationships between cotton leaf spectra curves (380-700 nm, 700-1300 nm, and 1300-2500 nm) and nitrogen content contributed to satisfactory predictions for nitrogen content detection. There are indeed many researches on the detection of LNC (Sun et al, 2013;Wang et al, 2018;Gao et al, 2022;Pourdarbani et al, 2022;Tang et al, 2022;Zhang et al, 2022b). Although the studies focusing on the LNC classification achieved good results (Sun et al, 2013;Wang et al, 2018;Pourdarbani et al, 2022), the samples in these studies were classified according to different nitrogen fertilization levels or different nitrogen fertilization days.…”