Nondestructive nitrogen content estimation in tomato plant leaves by Vis-NIR hyperspectral imaging and regression data models

A Multi-Target Regression Method to Predict Element Concentrations in Tomato Leaves Using Hyperspectral Imaging

Recent years have seen the development of novel, rapid, and inexpensive techniques for collecting plant data to monitor the nutritional status of crops. These techniques include hyperspectral imaging, which has been widely used in combination with machine learning models to predict element concentrations in plants. When there are multiple elements, the machine learning models are trained with spectral features to predict individual element concentrations; this type of single-target prediction is known as single-target regression. Although this method can achieve reliable accuracy for some elements, there are others that remain less accurate. We aimed to improve the accuracy of element concentration predictions by using a multi-target regression method that sequentially augmented the original input features (hyperspectral imaging) by chaining the predicted element concentration values. To evaluate the multi-target method, the concentrations of 17 elements in tomato leaves were predicted and compared with the single-target regression results. We trained 5 machine learning models with hyperspectral data and predicted element concentration values and found a significant improvement in the prediction accuracy for 10 elements (Mg, P, S, Mn, Fe, Co, Cu, Sr, Mo, and Cd). Furthermore, our multi-target regression method outperformed single-target predictions by increasing the coefficient of determination (R2) for elements such as Mn, Cu, Co, Fe, and Mg by 12.5%, 10.3%, 11%, 10%, and 8.4%, respectively. Hence, our multi-target method can improve the accuracy of predicting 10-element concentrations compared to single-target regression.

Rapid determination of the main components of corn based on near-infrared spectroscopy and a BiPLS-PCA-ELM model

To evaluate corn quality quickly, the feasibility of near-infrared spectroscopy (NIRS) coupled with chemometrics was analyzed to detect the moisture, oil, protein, and starch content in corn. A backward interval partial least squares (BiPLS)-principal component analysis (PCA)-extreme learning machine (ELM) quantitative analysis model was constructed based on BiPLS in conjunction with PCA and the ELM. The selection of characteristic spectral intervals was accomplished by BiPLS. The best principal components were determined by the prediction residual error sum of squares of Monte Carlo cross validation. In addition, a genetic simulated annealing algorithm was utilized to optimize the parameters of the ELM regression model. The established regression models for moisture, oil, protein, and starch can meet the demand for corn component detection with the prediction determination coefficients of 0.996, 0.990, 0.974, and 0.976; the prediction root means square errors of 0.018, 0.016, 0.067, and 0.109; and the residual prediction deviations of 15.704, 9.741, 6.330, and 6.236, respectively. The results show that the NIRS rapid detection model has higher robustness and accuracy based on the selection of characteristic spectral intervals in conjunction with spectral data dimensionality reduction and nonlinear modeling and can be used as an alternative strategy to detect multiple components in corn rapidly.

Nondestructive detection of anthocyanin content in fresh leaves of purple maize using hyperspectral data

Anthocyanins are widely used in the food industry as an additive, improving antioxidant capacity and strengthening the human immune system. However, rapid and nondestructive detection methods are lacking. This study aimed to develop a rapid and nondestructive method to detect anthocyanin content in fresh purple maize leaves using hyperspectral reflectance. Sensitivity bands were screened by analyzing the correlation between the spectrum and anthocyanin, chlorophyll, and moisture content in maize leaves with models constructed. Through a combination of the sensitivity bands of the three components, the interference of chlorophyll and moisture on the spectral detection of anthocyanin in fresh leaves was analyzed. The results showed that the anthocyanin sensitivity band was approximately 550 nm. The determination coefficient and root mean square error of the optimal hyperspectral model were 0.766 and 4.215 mg/g, respectively. After excluding chlorophyll and moisture interference, the anthocyanin content detection accuracy was improved by only 2% compared to that of the original. These results indicate that hyperspectral technology can be used to nondestructively detect anthocyanin content in fresh purple maize leaves with good accuracy. Chlorophyll and moisture in the leaves did not significantly influence anthocyanin content.

One-Dimensional Convolutional Neural Networks for Hyperspectral Analysis of Nitrogen in Plant Leaves

Accurately determining the nutritional status of plants can prevent many diseases caused by fertilizer disorders. Leaf analysis is one of the most used methods for this purpose. However, in order to get a more accurate result, disorders must be identified before symptoms appear. Therefore, this study aims to identify leaves with excessive nitrogen using one-dimensional convolutional neural networks (1D-CNN) on a dataset of spectral data using the Keras library. Seeds of cucumber were planted in several pots and, after growing the plants, they were divided into different classes of control (without excess nitrogen), N30% (excess application of nitrogen fertilizer by 30%), N60% (60% overdose), and N90% (90% overdose). Hyperspectral data of the samples in the 400–1100 nm range were captured using a hyperspectral camera. The actual amount of nitrogen for each leaf was measured using the Kjeldahl method. Since there were statistically significant differences between the classes, an individual prediction model was designed for each class based on the 1D-CNN algorithm. The main innovation of the present research resides in the application of separate prediction models for each class, and the design of the proposed 1D-CNN regression model. The results showed that the coefficient of determination and the mean squared error for the classes N30%, N60% and N90% were 0.962, 0.0005; 0.968, 0.0003; and 0.967, 0.0007, respectively. Therefore, the proposed method can be effectively used to detect over-application of nitrogen fertilizers in plants.