Search

Breeding high-photosynthetic-efficiency wheat varieties is a crucial link in safeguarding national food security. Traditional identification methods necessitate laborious on-site observation and measurement, consuming time and effort. Leveraging unmanned aerial vehicle (UAV) remote sensing technology to forecast photosynthetic indices opens up the potential for swiftly discerning high-photosynthetic-efficiency wheat varieties. The objective of this research is to develop a multi-stage predictive model encompassing nine photosynthetic indicators at the field scale for wheat breeding. These indices include soil and plant analyzer development (SPAD), leaf area index (LAI), net photosynthetic rate (Pn), transpiration rate (Tr), intercellular CO2 concentration (Ci), stomatal conductance (Gsw), photochemical quantum efficiency (PhiPS2), PSII reaction center excitation energy capture efficiency (Fv'/Fm'), and photochemical quenching coefficient (qP). The ultimate goal is to differentiate high-photosynthetic-efficiency wheat varieties through model-based predictions. This research gathered red, green, and blue spectrum (RGB) and multispectral (MS) images of eleven wheat varieties at the stages of jointing, heading, flowering, and filling. Vegetation indices (VIs) and texture features (TFs) were extracted as input variables. Three machine learning regression models (Support Vector Machine Regression (SVR), Random Forest (RF), and BP Neural Network (BPNN)) were employed to construct predictive models for nine photosynthetic indices across multiple growth stages. Furthermore, the research conducted principal component analysis (PCA) and membership function analysis on the predicted values of the optimal models for each indicator, established a comprehensive evaluation index for high photosynthetic efficiency, and employed cluster analysis to screen the test materials. The cluster analysis categorized the eleven varieties into three groups, with SH06144 and Yannong 188 demonstrating higher photosynthetic efficiency. The moderately efficient group comprises Liangxing 19, SH05604, SH06085, Chaomai 777, SH05292, Jimai 22, and Guigu 820, totaling seven varieties. Xinmai 916 and Jinong 114 fall into the category of lower photosynthetic efficiency, aligning closely with the results of the clustering analysis based on actual measurements. The findings suggest that employing UAV-based multi-source remote sensing technology to identify wheat varieties with high photosynthetic efficiency is feasible. The study results provide a theoretical basis for winter wheat phenotypic monitoring at the breeding field scale using UAV-based multi-source remote sensing, offering valuable insights for the advancement of smart breeding practices for high-photosynthetic-efficiency wheat varieties. [ABSTRACT FROM AUTHOR]

An efficient and accurate estimation of wheat growth and yield is important for wheat assessment and field management. To improve the accuracy and stability of wheat growth and yield estimation, an estimation method based on a genetic algorithm-improved support vector regression (GA-SVR) algorithm was proposed in this study. The correlation analysis between vegetation indices calculated from spectral data and wheat growth phenotypes and yields was performed to obtain the optimal combination of vegetation indices with high correlation and good estimation performance. At the same time, the optimal model for wheat growth monitoring was screened and constructed in experiments with 12 wheat varieties and 3 gradient nitrogen fertilizer application levels. Then, the yield estimation model was established and its applicability was verified under different nitrogen fertilizer application levels. The results showed that the constructed models for the leaf area index, plant height, and yield estimation performed well, with coefficients of determination of 0.82, 0.71, and 0.70, and root mean square errors of 0.09, 2.7, and 68.5, respectively. This study provided an effective UAV remote sensing technique for monitoring wheat growth status and estimating yield. This study provides an effective unmanned aerial remote sensing technique for monitoring wheat growth and estimating yield, and provides technical support for wheat yield assessment and field management. [ABSTRACT FROM AUTHOR]

The agriculture sector plays a very important role in the increasing population, year by year, to fulfill requirements and contributes significantly to the economy of the country. One of the main challenges in agriculture is the prevention and early detection of pest attack on crops. Farmers spend a significant amount of time and money in detecting pests and diseases, often by looking at plant leaves and analyzing the presence of diseases and pests. Late detection of pest attacks and improper use of pesticide application can cause damage to plants and compromise food quality. This problem can be solved through artificial intelligence, machine learning, and accurate image classification systems. In recent years, machine learning has made improvements in image recognition and classification. Hence, in this research article, we used convolutional neural network (CNN)-based models, such as the Cov2D library and VGG-16, to identify pest attacks. Our experiments involved a personal dataset consisting of 7000 images of pest-attacked leaf samples of different positions on maize plants, categorized into two classes. The Google Colab environment was used for experimentation and implementation, specially designed for cloud computing and machine learning. [ABSTRACT FROM AUTHOR]

The LAI (leaf area index) is an important parameter describing the canopy structure of citrus trees and characterizing plant photosynthesis, as well as providing an important basis for selecting parameters for orchard plant protection operations. By fusing LiDAR data with multispectral data, it can make up for the lack of rich spatial features of multispectral data, thus obtaining higher LAI inversion accuracy. This study proposed a multiscale LAI inversion method for citrus orchard based on the fusion of point cloud data and multispectral data. By comparing various machine learning algorithms, the mapping relationship between the characteristic parameters in multispectral data and point cloud data and citrus LAI was established, and we established the inversion model based on this, by removing redundant features through redundancy analysis. The experiment results showed that the BP neural network performs the best at both the community scale and the individual scale. After removing redundant features, the R2, RMSE, and MAE of the BP neural network at the community scale and individual scale were 0.896, 0.112, 0.086, and 0.794, 0.408, 0.328, respectively. By adding the three-dimensional gap fraction feature to the two-dimensional vegetation index features, the R2 at community scale and individual scale increased by 4.43% and 7.29%, respectively. The conclusion of this study suggests that the fusion of point cloud and multispectral data exhibits superior accuracy in multiscale citrus LAI inversion compared to relying solely on a single data source. This study proposes a fast and efficient multiscale LAI inversion method for citrus, which provides a new idea for the orchard precise management and the precision of plant protection operation. [ABSTRACT FROM AUTHOR]

In the present study, we used device visualization in tandem with deep learning to detect weeds in the wheat crop system in actual time. We selected the PMAS Arid Agriculture University research farm and wheat crop fields in diverse weather environments to collect the weed images. Some 6000 images were collected for the study. Throughout the season, tfhe databank was assembled to detect the weeds. For this study, we used two different frameworks, TensorFlow and PyTorch, to apply deep learning algorithms. PyTorch's implementation of deep learning algorithms performed comparatively better than that of TensorFlow. We concluded that the neural network implemented through the PyTorch framework achieves a superior outcome in speed and accuracy compared to other networks, such as YOLO variants. This work implemented deep learning models for weed detection using different frameworks. While working on real-time detection models, it is very important to consider the inference time and detection accuracy. Therefore, we have compared the results in terms of execution time and prediction accuracy. In particular, the accuracy of weed removal from wheat crops was judged to be 0.89 and 0.91, respectively, with inference times of 9.43 ms and 12.38 ms on the NVIDIA RTX2070 GPU for each picture (640 × 640). [ABSTRACT FROM AUTHOR]

Unknown objects in agricultural soil can be important because they may impact the health and productivity of the soil and the crops that grow in it. Challenges in collecting soil samples present opportunities to utilize Ground Penetrating Radar (GPR) image processing and artificial intelligence techniques to identify and locate unidentified objects in agricultural soil, which are important for agriculture. In this study, we used finite-difference time-domain (FDTD) simulated models to gather training data and predict actual soil conditions. Additionally, we propose a multi-class support vector machine (MSVM) that employs a semi-supervised algorithm to classify buried object materials and locate their position in soil. Then, we extract echo signals from the electromagnetic features of the FDTD simulation model, including soil type, parabolic shape, location, and energy magnitude changes. Lastly, we compare the performance of various MSVM models with different kernel functions (linear, polynomial, and radial basis function). The results indicate that the FDTD-Yee method enhances the accuracy of simulating real agricultural soils. The average recognition rate of the hyperbola position formed by the GPR echo signal is 91.13%, which can be utilized to detect the position and material of unknown and underground objects. For material identification, the directed acyclic graph support vector machine (DAG-SVM) model attains the highest classification accuracy among all soil layers when using an RBF kernel. Overall, our study demonstrates that an artificial intelligence model trained with the FDTD forward simulation model can effectively detect objects in farmland soil. [ABSTRACT FROM AUTHOR]

As an important factor determining the competitiveness of raw cotton, cotton fiber quality has received more and more attention. The results of traditional detection methods are accurate, but the sampling cost is high and has a hysteresis, which makes it difficult to measure cotton fiber quality parameters in real time and at a large scale. The purpose of this study is to use time-series UAV (Unmanned Aerial Vehicle) multispectral and RGB remote sensing images combined with machine learning to model four main quality indicators of cotton fibers. A deep learning algorithm is used to identify and extract cotton boll pixels in remote sensing images and improve the accuracy of quantitative extraction of spectral features. In order to simplify the input parameters of the model, the stepwise sensitivity analysis method is used to eliminate redundant variables and obtain the optimal input feature set. The results of this study show that the R2 of the prediction model established by a neural network is improved by 29.67% compared with the model established by linear regression. When the spectral index is calculated after removing the soil pixels used for prediction, R2 is improved by 4.01% compared with the ordinary method. The prediction model can well predict the average length, uniformity index, and micronaire value of the upper half. R2 is 0.8250, 0.8014, and 0.7722, respectively. This study provides a method to predict the cotton fiber quality in a large area without manual sampling, which provides a new idea for variety breeding and commercial decision-making in the cotton industry. [ABSTRACT FROM AUTHOR]

Estimating plant physiological indicators with remote sensing technology is critical for ensuring precise field management. Compared with other remote sensing platforms, low-altitude unmanned aerial vehicles (UAVs) produce images with high spatial resolution that can be used to clearly identify vegetation. However, the information of UAV image data is relatively complex and difficult to analyze, which is the main problem limiting its large-scale use at present. In order to monitor plant physiological indexes from the multi-spectral data, a new method based on machine learning is studied in this paper. Using UAV for deriving the absorption coefficients of plant canopies and whole leaf area, this paper quantifies the effects of plant physiological indicators such as the soil and plant analyzer development (SPAD) value, whole leaf area, and dry matter accumulation on the relationship between the reflectance spectra. Nine vegetation indexes were then extracted as the sensitive vegetation indexes of the rice physiological indicators. Using the SVM model to predict the SPAD value of the plant, the mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and symmetric mean absolute percentage error (SMAPE) values of the model were 1.90, 1.38, 0.13, 0.86, and 4.13, respectively. The results demonstrate that the rice plants display a considerable biochemical and spectral correlation. Using SVM to predict the SPAD value has a better effect because of a better adaptation and a higher accuracy than other models. This study suggests that the multi-spectral data acquired using UAV can quickly estimate field physiological indicators, which has potential in the pre-visual detection of SPAD value information in the field. At the same time, it can also be extended to the detection and inversion of other key variables of crops. [ABSTRACT FROM AUTHOR]

The high-temporal-resolution monitoring of key management nodes in cotton management via agricultural remote sensing is vital for field cotton macro-statistics, particularly for predicting cotton production and obtaining comprehensive data. This study examines Shihezi, Xinjiang as a case study, utilizing Sentinel-1 and Sentinel-2 data from 2019 to 2021. Three machine learning models(RF, SVM, and CART) were employed to extract annual crop classification area rasters, monitor weekly cultivation progress, and monitor abandoned cropland during the cultivation period. The results demonstrate that the random forest model has produced satisfactory results in gridded extraction for cotton classification areas, achieving the producer's accuracy of the cotton category reached 98.5%, and the kappa coefficient is 0.947. Cotton cultivated in 2021 began is a week later than in 2020, yet exhibited a faster cultivate speed. The proportion of abandoned cotton fields in the study area rose in 2020 compared to 2019. The methodology presented in this study has a certain reference value for exploring the monitoring of continuous changes in crops over the years and macro-monitoring of important activities in the entire growth cycle. [ABSTRACT FROM AUTHOR]

Effective detection of rice spikelet flowering is crucial to the determination of optimal pollination timing for hybrid rice seed production. Currently, the detection of rice spikelet flowering status relies on manual observation of farmers, which has low efficiency and large errors. This study attempts to acquire rice spikelet flowering information using a hyperspectral technique and machine learning in order to meet the needs of hybrid rice seed pollination rapidly and automatically. Hyperspectral data of rice male parents with flowering and non-flowering in two experimental sites were collected with an ASD FieldSpec® HandHeld™2 spectrometer. Three traditional classifiers, Random Forest (RF), Support Vector Machine (SVM) and Back Propagation (BP) neural network, and Convolutional Neural Network (CNN), were used to build classification models for rice spikelets flowering detection. Three data processing methods, PCA feature extraction, GA feature selection, and the PCA and GA combination algorithm, were used for data dimensionality reduction. By comparing the precision and recall rate of different algorithms and data processing methods, the algorithms applicable to identify rice spikelet flowering were investigated. Results show that by evaluating different feature reduction methods and classifiers, the optimal model for rice spikelets flowering detection is the BP model with PCA feature extraction. The accuracy of the model reaches up to 96–100%. Hyperspectral technology and machine learning algorithm are capable of effective detection of rice spikelet flowering. This study provides technical reference for accurate judgment of rice flowering and helps to determine the optimal operation time for supplementary pollination of hybrid rice. [ABSTRACT FROM AUTHOR]

Citrus Huanglongbing (HLB), also named citrus greening disease, occurs worldwide and is known as a citrus cancer without an effective treatment. The symptoms of HLB are similar to those of nutritional deficiency or other disease. The methods based on single-source information, such as RGB images or hyperspectral data, are not able to achieve great detection performance. In this study, a multi-modal feature fusion network, combining a RGB image network and hyperspectral band extraction network, was proposed to recognize HLB from four categories (HLB, suspected HLB, Zn-deficient, and healthy). Three contributions including a dimension-reduction scheme for hyperspectral data based on a soft attention mechanism, a feature fusion proposal based on a bilinear fusion method, and auxiliary classifiers to extract more useful information are introduced in this manuscript. The multi-modal feature fusion network can effectively classify the above four types of citrus leaves and is better than single-modal classifiers. In experiments, the highest accuracy of multi-modal network recognition was 97.89% when the amount of data was not very abundant (1,325 images of the four aforementioned types and 1,325 pieces of hyperspectral data), while the single-modal network with RGB images only achieved 87.98% recognition and the single-modal network using hyperspectral information only 89%. Results show that the proposed multi-modal network implementing the concept of multi-source information fusion provides a better way to detect citrus HLB and citrus deficiency. [ABSTRACT FROM AUTHOR]

Spider mites are important pests that cause severe economic damage to cotton. They feed on underside of leaves, piercing the chloroplast-containing cells, resulting in foliar damage and yield reduction. This paper proposed a two-stage classification approach for mite-infestation detection based on machine learning methods. Two cotton fields were selected for study, and the UAV imagery collection and concurrent ground investigation were conducted on July 20-21th, 2017. Mosaicking and geo-registration were performed on the collected multispectral imagery. Support Vector Machine (SVM) was used for scene classification, and a transferred Convolutional Neural Network (CNN) was applied for mite-infestation identification. Experimental results showed that our method outperformed others in terms of accuracy, which demonstrated that our approach has potential in mite-infestation detection using UAV multispectral imagery. [ABSTRACT FROM AUTHOR]

Appropriate Site Specific Weed Management (SSWM) is crucial to ensure the crop yields. Within SSWM of large-scale area, remote sensing is a key technology to provide accurate weed distribution information. Compared with satellite and piloted aircraft remote sensing, unmanned aerial vehicle (UAV) is capable of capturing high spatial resolution imagery, which will provide more detailed information for weed mapping. The objective of this paper is to generate an accurate weed cover map based on UAV imagery. The UAV RGB imagery was collected in 2017 October over the rice field located in South China. The Fully Convolutional Network (FCN) method was proposed for weed mapping of the collected imagery. Transfer learning was used to improve generalization capability, and skip architecture was applied to increase the prediction accuracy. After that, the performance of FCN architecture was compared with Patch_based CNN algorithm and Pixel_based CNN method. Experimental results showed that our FCN method outperformed others, both in terms of accuracy and efficiency. The overall accuracy of the FCN approach was up to 0.935 and the accuracy for weed recognition was 0.883, which means that this algorithm is capable of generating accurate weed cover maps for the evaluated UAV imagery. [ABSTRACT FROM AUTHOR]

• A HLB field detection method based on multispectral images collected by UAV was investigated. • To explore a real-time detection method to recognize Citrus HLB in the air. • Different feature extraction methods and machine learning algorithms were compared to build a HLB detection model. Citrus Huanglongbing (HLB), also known as citrus greening, is the most destructive disease in the citrus industry. Detecting this disease as early as possible and eradicating the roots of HLB-infected trees can control its spread. Ground diagnosis is time-consuming and laborious. Large area monitoring method of citrus orchard with high accuracy is rare. This study evaluates the feasibility of large area detection of citrus HLB by low altitude remote sensing and commits to improve the accuracy of large-area detection. A commercial multispectral camera (ADC-lite) mounted on DJI M100 UAV(unmanned Aerial Vehicle) was used to collect green, red and near-infrared multispectral image of large area citrus orchard, a linear-stretch was performed to remove noise pixel, vegetation indices (VIs) were calculated followed by correlation analysis and feature compression using PCA (principal components analysis) and AutoEncoder to discover potential features. Several machine learning algorithms, such as support vector machine (SVM), k-nearest neighbour (kNN), logistic regression (LR), naive Bayes and ensemble learning, were compared to model the healthy and HLB-infected samples after parameter optimization. The results showed that the feature of PCA features of VIs combining with original DN (digital numbers) value generally have highest accuracy and agreement in all models, and the ensemble learning and neural network approaches had strong robustness and the best classification results (100% in AdaBoost and 97.28% in neural network) using threshold strategy. [ABSTRACT FROM AUTHOR]

• The purpose of this paper is to achieve the appearance quality inspection of citrus, addressing the challenges of defect target and fruit morphology detection. • A detection algorithm for defect targets on citrus peel and a method for fruit morphology detection have been proposed. • A citrus quality detection data set is constructed and evaluated. Experimental results show that the method in this paper achieves satisfactory results in performance. Identifying defects in citrus peels and analyzing fruit morphology are two core challenges in citrus quality inspection. In order to more accurately identify minor defects on citrus peels, we proposed a detection model Yolo-FD (Yolo for defects). The model was based on the Yolov5 network framework, and the backbone network embedded the Three-dimensional Coordinate Attention (TDCA) mechanism innovatively designed in this study. It accurately captured the subtle changes and feature associations of the target in spatial location, significantly enhancing the model's ability to perceive defects in fruit peels. Moreover, we employed a simplified Bidirectional Weighted Feature Pyramid Network (BiFPN) in the model to achieve cross-scale connections and improve the feature fusion ability of the model. At the same time, Contextual Transformer block (COT) was introduced into Neck network and the CoT3 module was built to fully capture the static and dynamic contextual information in the citrus defects images and enhance the expression of the feature map. Through this series of improvement methods, missed detections and false detections caused by small targets were effectively reduced. Fruit morphology detection was combined with the Partice Swarm Optimized Extreme Learning Machine (PSO-ELM) model to determine whether the citrus fruit morphology was well-formed, using the symmetry index, roundness and tilt angle of the citrus as input parameters. The experimental results indicated that the mean average precision of the Yolo-FD model is 98.7 % (mAP-0.5). Compared with Yolov5s, Yolov7-tiny, and Yolov8n, the mAP was improved by 1.4 %, 1.5 %, and 0.5 % respectively. Its average detection time for a single frame image on the server was 19.5 ms. And the PSO-ELM model achieved a fruit morphology detection accuracy of 91.42 %, a coefficient of determination of 0.9044, and a mean squared error of 0.8497. The research results met the accuracy and real-time requirements for citrus sorting on the production line, and could provide an effective solution for citrus grading and quality assessment. [ABSTRACT FROM AUTHOR]

• The PLL/WS2/LIG electrode with 2D-on-2D layered structure exhibits an excellent electrochemical property towards CBZ. • The flexible electrochemical sensor, encapsulated by 3D printing, achieve the non-destructive in-situ detection of CBZ on tea leaves. • Easy-to-use visual online cloud platform is developed for convenient and intelligent detection. • Real-time intelligent analysis is realized to monitor the change in CBZ residues to guarantee food safety. Excessive use of carbendazim (CBZ) for controlling fungal disease may result in pesticide residues in growth environments. There is an urgent requirement for real-time detection of CBZ content to enable rapid assessment of pesticide residues. Herein, a novel electrochemical sensor was created, which utilizes a laser-induced graphene (LIG) electrode modified with two-dimensional (2D) tungsten disulfide (WS 2) nanosheets and a poly(L-lysine) (PLL) film to construct a 2D-on-2D layered structure. The PLL/WS 2 /LIG electrode exhibits strong charge transfer, excellent electrocatalytic performance, and high CBZ adsorption capacity, thus leading to enhanced sensitivity. As a result, the sensor exhibits a low detection limit of 1.2 ng/mL (6.2 nM), high sensitivity up to 135.1 μA(μg/mL)−1cm−2, and a wide response range of 0.01 to 1.5 μg/mL. Moreover, an intelligent analysis system, consisting of a portable device that connects to an online cloud platform equipped with a machine learning model and visualization software, has been developed for the real-time detection of CBZ residues on tea leaves in tea plantations. This study demonstrates that the PLL/WS 2 /LIG sensor and the intelligent analysis system have great potential to monitor pesticide residues to guarantee food safety. [Display omitted] [ABSTRACT FROM AUTHOR]

[Display omitted] • HSI with ML technique effectively predicted three pigment levels in tomato seedlings. • Feature extraction algorithms enhanced models by filtering out sensitive wavelengths. • The ELM model surpassed the PLSR model in prediction tasks.. • Classification models identified gene type of the seedlings based on pigment levels. Conventional chemical approaches could be limited in monitoring the concentration of pigments in plants in high volumes. To overcome these limitations, researchers often turn to non-invasive, high-throughput, and real-time monitoring techniques, such as spectroscopy and hyperspectral imaging, which allow for the assessment of pigment concentration in plants without the need for destructive sampling and offer the ability to monitor large volumes of plants efficiently. This research focused on the utilization of machine learning in conjunction with hyperspectral imaging to develop models for predicting the concentration of three pigments, namely chlorophyll-a, chlorophyll-b, and carotenoids, in tomato seedlings. The sample tomato seedlings were sourced from two distinct varieties: the wild type and the Long Hypocotyl-5-deficient (HY5) type. The spectral data were acquired using a near-infrared (NIR) camera with a spectral range spanning approximately 900–1700 nm. Machine learning algorithms such as partial least squares regression (PLSR) and extreme learning machine (ELM) were utilized to explore the latent relationship between hyperspectral information and chemical measurements. In addition, principal component analysis (PCA), independent component analysis (ICA), and competitive adaptive reweighted sampling (CARS) methods were used to extracted informative wavelengths from the reflectance spectrum. And a comprehensive analysis regarding to spectroscopy was conducted to investigate the validity and efficiency of the results of feature extraction. The ELM model demonstrated the highest effectiveness, achieving R2 values of 0.86, 0.83, and 0.83 for chlorophyll-a, chlorophyll-b, and carotenoids, respectively, on the test set. By integrating the predictive models with classifiers such as Logistic Regression, Support Vector Classifier (SVC), and K-nearest Neighbors (KNN), tomato seedlings were categorized into wild type and HY5 type. The findings showed that the proposed approach efficiently predicted the concentration of the pigments in tomato seedlings and explored the feasibility of using these results to identify tomato gene types. [ABSTRACT FROM AUTHOR]

• A method to create a defoliation spray map using predictive FPAR was devised. • GBDT is the best model for estimating the FPAR of cotton. • Fusion of multi-source remote sensing data improves model prediction accuracy. • Textural features outperform spectral indices in estimating cotton's FPAR. Efficient and accurate spraying of cotton defoliant is a vital part of cotton production; the traditional way of spraying cotton defoliant will cause waste of pesticides and environmental pollution. The fraction of absorbed photosynthetically active radiation(FPAR) in cotton predicted by multi-source remote sensing information from unmanned aerial vehicles(UAVs) offers the possibility of precise control of cotton defoliant dosage. In this study, RGB and multispectral(MS) images of cotton at multiple fertility stages were collected by UAV. Six machine learning models (random forest regression(RFR), support vector regression(SVR), gradient-boosting decision tree(GBDT), extreme gradient boosting(XGBoost), K-nearest neighbor(k-NN), partial least squares regression(PLSR)) were constructed using spectral indices and texture features extracted from UAV remote sensing imagery and cotton FPAR measurements collected, and the performance of each model was evaluated. A spatial distribution map of cotton FPAR based on UAV imagery was constructed from cotton FPAR predicted by the machine learning models. The study results show that (i) image feature changes over multiple periods based on UAV multi-source remote sensing data are consistent with the evolution of cotton FPAR and can be used to predict cotton FPAR. (ii) Overall, texture features are more sensitive to FPAR than spectral indices, and the fusion of texture features with spectral indices increases the prediction accuracy of the model. (iii) The GBDT model based on UAV multi-source remote sensing data not only performs better on cotton FPAR prediction for a large sample dataset (whole fertility period), with R2 = 0.88 and rRMSE = 7.43 %, but also for a small sample dataset (single fertility period), with a range of R2 of 0.65–0.85 and a range of rRMSE of 7.41 %-19.09 %. In addition, a cotton defoliation spray prescription map was constructed based on the spatial distribution of cotton FPAR predicted by GBDT. [ABSTRACT FROM AUTHOR]

As the most important and widely used spice in the world, black pepper is known as the "king of spices." The geographical origin of black pepper greatly affects its quality and price. The existing physicochemical detection methods for distinguishing black pepper have inherent performance issues, such as expensive equipment, complex operations and high time consumption levels. This study proposes a novel method for identifying the origin of black pepper by synergically applying an E-tongue (ET), an E-nose (EN) and an E-eye (EE) in combination with a deep learning algorithm. First, taste and smell fingerprints were collected by ET and EN instruments, respectively, and the color, shape and texture information of different samples was collected by EE instruments. Three kinds of convolutional neural networks (CNNs) with one-dimensional or two-dimensional convolutional structures were designed and utilized to extract the feature information from the ET, EN and EE signals. Additionally, the Bayesian optimization algorithm (BOA) was applied to globally optimize the hyperparameters of the different CNN models. Then, a channel attention mechanism (CAM) module was introduced to achieve feature-level fusion for the three kinds of signals. Finally, a fully connected layer that uses a softmax algorithm was utilized for classifying the categories of black pepper. The experimental results showed that compared with employing a single sensory device, the proposed method yielded better recognition accuracy. Achieving accuracy, precision, recall and F 1 -score values of 99.71%, 0.997, 0.997 and 0.996 respectively, the proposed pattern recognition model obtained better classification results than the baseline models for the test set. This study introduces a rapid detection method for identifying the geographical origin of black pepper. [Display omitted] • A synergetic application of ET, EN and EE systems combined with deep learning is explored. • 1D and 2D CNN models are designed to process different signal of intelligent senses. • A CAM is applied to fuse the features of the ET, EN and EE signals. • A low-cost and accurate detection technique is proposed. [ABSTRACT FROM AUTHOR]

[Display omitted] • Electronic tongue and eye are used to classify Pu-erh tea via feature-level fusion. • The deep learning is introduced to data fusion field of intelligent sensory system. • CNN structures are designed and optimized by Bayesian optimization. • The proposed model has better performance compared with conventional methods. This study proposed an efficient approach that an electronic tongue (ET) and an electronic eye (EE) combined with a deep learning algorithm were jointly leveraged to recognition Pu-erh tea. A one-dimensional convolutional neural network (1-D CNN) and a two-dimensional convolutional neural network (2-D CNN) were designed and optimized for the feature extraction of ET and EE signals, respectively. Then, a feature-level fusion strategy was introduced to address the feature vectors extracted from the different types of CNN models. To highlight the effect of data fusion, a backpropagation neural network (BPNN), a classifier similar to the fully connected layers of CNN models, was employed. Meanwhile, the Bayesian optimization algorithm (BOA) was employed for hyperparameter optimization of the identification models. The experimental results showed that the feature fusion strategy assimilated the merits of the ET and EE and gained better Pu-erh tea identification performance than an independent intelligent sensory system combined with CNN model. The results demonstrate that the feature-level fusion based on deep learning algorithm gained the best accuracy on the test set, with a precision, a recall, an F1-score and an AUC score of 99.07%, 99.2%, 0.992 and 0.994, respectively. This study shows that the simultaneous utilization of an ET and an EE combined with deep learning algorithm could function as a rapid detection method for discriminating the storage time of Pu-erh tea. [ABSTRACT FROM AUTHOR]

Pu-erh tea is a famous Chinese fermented tea, and its quality and flavor are closely related to the storage time used for its fermentation. This paper puts forward one method to discriminate the age of Pu-erh tea by employing a voltammetric electronic tongue (VE-Tongue) combined with deep learning and transfer learning techniques. To make the deep learning model suitable for processing VE-Tongue signals, a one-dimensional convolutional neural network (1-D CNN) was developed to automatically perform feature extraction and classification. Transfer learning (TL) was introduced to train the model to reduce the training complexity and enhance the generalization capability of the CNN. The performance of the proposed model was further compared with that of traditional machine learning methods such as the backpropagation neural network, support vector machine and extreme learning machine. The results showed that the proposed model exhibited better performance in classifying Pu-erh tea than other methods. Its accuracy for the test set, precision, recall and F1 score was 98.80%, 98.2%, 98%, and 0.98, respectively. This study found that the VE-Tongue combined with deep learning and TL algorithms could be a sensitive, reliable and effective detection method for identifying the amount of storage time of Pu-erh tea, which could further expand its applications to other related fields. Image 1 • Classification of Pu-erh tea is achieved by voltammetric electronic tongue system. • Pattern recognition method based on deep learning and transfer learning is proposed. • 1-D CNN model is design and optimized for analyzing electronic tongue signals. • The proposed model is superior to conventional pattern recognition methods. [ABSTRACT FROM AUTHOR]

• Authentication of honey form its botanical origin is achieved by electronic tongue. • A novel remote electronic tongue system is developed and applied. • A new feature extraction method is used for analyzing electronic tongue signals. • The approach may be used as an on-site tool for honey authentication. The misrepresentation of honey botanical origin causes serious economic losses for honey producers and consumers. In this work, a novel remote electronic tongue system was developed and utilized to identify honey from different botanical origins. The system mainly comprises an electrode array, a portable instrument, a general packet radio service (GPRS) transmitter module, and a remote server. To reduce the number of redundant variables, a feature extraction method combining variational mode decomposition and Hilbert transformation (VMD-HT) is proposed. Then, the classification algorithm, principal component analysis (PCA) and extreme learning machine (ELM) are applied to discriminate different honey samples. The experiments show that the proposed method obtains more accurate classification results than other existing feature extraction methods. The developed electronic tongue system provides a rapid, portable and cost-effective method for honey botanical origin authentication, which can be effectively applied in the honey industry. [ABSTRACT FROM AUTHOR]

• A field detection method of citrus Huanglongbing based on hyperspectral data was investigated. • To recognize asymptomatic leaves from healthy one so that explore an early detection method. • The band selection method based on entropy distance and sequence backward selection was developed. • Different feature extraction and machine learners were compared to model the different groups of the disease. Citrus Huanglongbing (HLB), also called citrus greening, is the most destructive disease in the citrus industry. Detecting the disease as early as possible and then eradicating infected roots can effectively control its spread. For the Shatangju mandarin cultivar, a non-destructive citrus HLB field detection method based on hyperspectral reflectance is proposed in this study. A characteristic band extraction method based on entropy distance and sequential backward selection is explored. Several machine learning algorithms (logistic regression, decision tree, support vector machine, k-nearest neighbor, linear discriminant analysis, and ensemble learning) were used to discriminate between disease groups: healthy, symptomatic HLB-infected, and asymptomatic HLB-infected, based on leaf reflectance. The results showed that the use of primary hyperspectral reflectance is very feasible for such classification. The band selection method proposed in this study provides an option for dimensionality reduction while still providing high classification accuracy. In three-group classification, the SVM learner achieved 90.8% accuracy, while in two-group classification (healthy vs symptomatic HLB leaves), the accuracy reached to 96%. The results also show that using only a few bands is insufficient for classification. In this study, 13 characteristic bands extracted by the proposed method provided the best performance. [ABSTRACT FROM AUTHOR]