Search

Abstract Currently, image recognition based on deep neural networks has become the mainstream direction of research; therefore, significant progress has been made in its application in the field of tea detection. Many deep models exhibit high recognition rates in tea leaves detection. However, deploying these models directly on tea-picking equipment in natural environments is impractical; the extremely high parameters and computational complexity of these models make it challenging to perform real-time tea leaves detection. Meanwhile, lightweight models struggle to achieve competitive detection accuracy; therefore, this paper addresses the issue of computational resource constraints in remote mountain areas and proposes Reconstructed Feature and Dual Distillation (RFDD) to enhance the detection capability of lightweight models for tea leaves. In our method, the Reconstructed Feature selectively masks the feature of the student model based on the spatial attention map of the teacher model; it utilizes a generation block to force the student model to generate the teacher’s full feature. The Dual Distillation comprises Decoupled Distillation and Global Distillation. Decoupled Distillation divides the reconstructed feature into foreground and background features based on the Ground-Truth. This compels the student model to allocate different attention to foreground and background, focusing on their critical pixels and channels. However, Decoupled Distillation leads to the loss of relation knowledge between foreground and background pixels. Therefore, we further perform Global Distillation to extract this lost knowledge. Since RFDD only requires loss calculation on feature map, it can be easily applied to various detectors. We conducted experiments on detectors with different frameworks, using a tea dataset collected at the Huangshan Houkui Tea Plantation. The experimental results indicate that, under the guidance of RFDD, the student detectors have achieved performance improvements to varying degrees. For instance, a one-stage detector like RetinaNet (ResNet-50) experienced a 3.14% increase in Average Precision (AP) after RFDD guidance. Similarly, a two-stage model like Faster RCNN (ResNet-50) obtained a 3.53% improvement in AP. This offers promising prospects for lightweight models to efficiently perform real-time tea leaves detection tasks.

Abstract Low‐light image enhancement is a longstanding challenge in low‐level vision, as images captured in low‐light conditions often suffer from significant aesthetic quality flaws. Recent methods based on deep neural networks have made impressive progress in this area. In contrast to mainstream convolutional neural network (CNN)‐based methods, an effective solution inspired by the transformer, which has shown impressive performance in various tasks, is proposed. This solution is centred around two key components. The first is an image synthesis pipeline, and the second is a powerful transformer‐based pre‐trained model, known as the low‐light image enhancement transformer (LIET). The image synthesis pipeline includes illumination simulation and realistic noise simulation, enabling the generation of more life‐like low‐light images to overcome the issue of data scarcity. LIET combines streamlined CNN‐based encoder‐decoders with a transformer body, efficiently extracting global and local contextual features at a relatively low computational cost. The extensive experiments show that this approach is highly competitive with current state‐of‐the‐art methods. The codes have been released and are available at LIET.

Abstract Non-coding RNAs (ncRNAs) are recognized as pivotal players in the regulation of essential physiological processes such as nutrient homeostasis, development, and stress responses in plants. Common methods for predicting ncRNAs are susceptible to significant effects of experimental conditions and computational methods, resulting in the need for significant investment of time and resources. Therefore, we constructed an ncRNA predictor(MFPINC), to predict potential ncRNA in plants which is based on the PINC tool proposed by our previous studies. Specifically, sequence features were carefully refined using variance thresholding and F-test methods, while deep features were extracted and feature fusion were performed by applying the GRU model. The comprehensive evaluation of multiple standard datasets shows that MFPINC not only achieves more comprehensive and accurate identification of gene sequences, but also significantly improves the expressive and generalization performance of the model, and MFPINC significantly outperforms the existing competing methods in ncRNA identification. In addition, it is worth mentioning that our tool can also be found on Github ( https://github.com/Zhenj-Nie/MFPINC ) the data and source code can also be downloaded for free.

Abstract Knowledge distillation can transfer the knowledge from the pre-trained teacher model to the student model, thus effectively accomplishing model compression. Previous studies have carefully crafted knowledge representation, targeting loss function design, and distillation location selection, but there have been few studies on the role of classifiers in distillation. Previous experiences have shown that the final classifier of the model has an essential role in making inferences, so this paper attempts to narrow the gap in performance between models by having the student model directly use the classifier of the teacher model for the final inference, which requires an additional projector to help match features of the student encoder with the teacher's classifier. However, a single projector cannot fully align the features, and integrating multiple projectors may result in better performance. Considering the balance between projector size and performance, through experiments, we obtain the size of projectors for different network combinations and propose a simple method for projector integration. In this way, the student model undergoes feature projection and then uses the classifiers of the teacher model for inference, obtaining a similar performance to the teacher model. Through extensive experiments on the CIFAR-100 and Tiny-ImageNet datasets, we show that our approach applies to various teacher–student frameworks simply and effectively.

Bruising reduces the edibility and marketability of fresh apples, inevitably causing economic losses for the apple industry. However, bruises lack obvious visual symptoms, which makes it challenging to detect them using imaging techniques with uniform or diffuse illumination. This study employed the structured light imaging (SLI) technique to detect apple bruises. First, the grayscale reflection images were captured under phase-shifted sinusoidal illumination at three different wavelengths (600, 650, and 700 nm) and six different spatial frequencies (0.05, 0.10, 0.15, 0.20, 0.25, and 0.30 cycles mm−1). Next, the grayscale reflectance images were demodulated to produce direct component (DC) images representing uniform diffuse illumination and amplitude component (AC) images revealing bruises. Then, by quantifying the contrast between bruised regions and sound regions in all AC images, it was found that bruises exhibited the optimal contrast when subjected to sinusoidal illumination at a wavelength of 700 nm and a spatial frequency of 0.25 mm−1. In the AC image with optimal contrast, the developed h-domes segmentation algorithm to accurately segment the location and range of the bruised regions. Moreover, the algorithm successfully accomplished the task of segmenting central bruised regions while addressing the challenge of segmenting edge bruised regions complicated by vignetting. The average Intersection over Union (IoU) values for the three types of bruises were 0.9422, 0.9231, and 0.9183, respectively. This result demonstrated that the combination of SLI and the h-domes segmentation algorithm was a viable approach for the effective detection of fresh apple bruises.

Abstract Background More and more studies show that miRNA plays a crucial role in plants' response to different abiotic stresses. However, traditional experimental methods are often expensive and inefficient, so it is important to develop efficient and economical computational methods. Although researchers have developed machine learning-based method, the information of miRNAs and abiotic stresses has not been fully exploited. Therefore, we propose a novel approach based on graph neural networks for predicting potential miRNA-abiotic stress associations. Results In this study, we fully considered the multi-source feature information from miRNAs and abiotic stresses, and calculated and integrated the similarity network of miRNA and abiotic stress from different feature perspectives using multiple similarity measures. Then, the above multi-source similarity network and association information between miRNAs and abiotic stresses are effectively fused through heterogeneous networks. Subsequently, the Restart Random Walk (RWR) algorithm is employed to extract global structural information from heterogeneous networks, providing feature vectors for miRNA and abiotic stress. After that, we utilized the graph autoencoder based on GIN (Graph Isomorphism Networks) to learn and reconstruct a miRNA-abiotic stress association matrix to obtain potential miRNA-abiotic stress associations. The experimental results show that our model is superior to all known methods in predicting potential miRNA-abiotic stress associations, and the AUPR and AUC metrics of our model achieve 98.24% and 97.43%, respectively, under five-fold cross-validation. Conclusions The robustness and effectiveness of our proposed model position it as a valuable approach for advancing the field of miRNA-abiotic stress association prediction.

Visual navigation is the pivotal technology for enabling autonomous operations of orchard robots. To obtain orchard navigation lines, the robot needs to quickly identify the positions of tree trunks. For this, we proposed a detection model called YOLOv8-Trunk in this study. Based on the detection results of vine tree trunks by YOLOv8-Trunk, the network generates a series of center point coordinates at the bottom of the detection boxes. Subsequently, the least square method is employed to fit reference lines on both sides of the trunk, thereby determining the navigation path for the orchard robot. To enhance the focus on the target, an efficient multi-scale attention (EMA) mechanism is introduced into traditional YOLOv8 network. On the data level, we adopted a novel Mix-Shelter method to augment the datasets for training the detection model, thereby bolstering the robustness. In addition, we also explored the impact of loss functions and optimizers on the performance of the detection model. A comprehensive set of ablation and comparison experiments is conducted in this study. The experimental results affirm that the YOLOv8-Trunk network adeptly detects vine tree trunks, achieving an accuracy rate of 92.7%. The obtained navigation path based on the detect result is reliable. This study provides valuable reference for the realization of intelligent inspection in orchards.

To improve detection efficiency and reduce cost consumption in fishery surveys, target detection methods based on computer vision have become a new method for fishery resource surveys. However, the specialty and complexity of underwater photography result in low detection accuracy, limiting its use in fishery resource surveys. To solve these problems, this study proposed an accurate method named BSSFISH-YOLOv8 for fish detection in natural underwater environments. First, replacing the original convolutional module with the SPD-Conv module allows the model to lose less fine-grained information. Next, the backbone network is supplemented with a dynamic sparse attention technique, BiFormer, which enhances the model’s attention to crucial information in the input features while also optimizing detection efficiency. Finally, adding a 160 × 160 small target detection layer (STDL) improves sensitivity for smaller targets. The model scored 88.3% and 58.3% in the two indicators of mAP@50 and mAP@50:95, respectively, which is 2.0% and 3.3% higher than the YOLOv8n model. The results of this research can be applied to fishery resource surveys, reducing measurement costs, improving detection efficiency, and bringing environmental and economic benefits.

Vision-based underwater exploration is crucial for marine research. However, the degradation of underwater images due to light attenuation and scattering poses a significant challenge. This results in the poor visual quality of underwater images and impedes the development of vision-based underwater exploration systems. Recent popular learning-based Underwater Image Enhancement (UIE) methods address this challenge by training enhancement networks with annotated image pairs, where the label image is manually selected from the reference images of existing UIE methods since the groundtruth of underwater images do not exist. Nevertheless, these methods encounter uncertainty issues stemming from ambiguous multiple-candidate references. Moreover, they often suffer from local perception and color perception limitations, which hinder the effective mitigation of wide-range underwater degradation. This paper proposes a novel NUAM-Net (Novel Underwater Image Enhancement Attention Mechanism Network) that addresses these limitations. NUAM-Net leverages a probabilistic training framework, measuring enhancement uncertainty to learn the UIE mapping from a set of ambiguous reference images. By extracting features from both the RGB and LAB color spaces, our method fully exploits the fine-grained color degradation clues of underwater images. Additionally, we enhance underwater feature extraction by incorporating a novel Adaptive Underwater Image Enhancement Module (AUEM) that incorporates both local and long-range receptive fields. Experimental results on the well-known UIEBD benchmark demonstrate that our method significantly outperforms popular UIE methods in terms of PSNR while maintaining a favorable Mean Opinion Score. The ablation study also validates the effectiveness of our proposed method.

The accurate identification, detection, and segmentation of tea buds and leaves are important factors for realizing intelligent tea picking. A tea picking point location method based on the region-based convolutional neural network(R-CNN) Mask- RCNN is proposed, and a tea bud and leaf and picking point recognition model is established. First, tea buds and leaf pictures are collected in a complex environment, the Resnet50 residual network and a feature pyramid network (FPN) are used to extract bud and leaf features, and preliminary classification and preselection box regression training-performed on the feature maps through a regional proposal network (RPN). Second, the regional feature aggregation method (RoIAlign) is used to eliminate the quantization error, and the feature map of the preselected region of interest (ROI) is converted into a fixed-size feature map. The output module of the model realizes the functions of classification, regression and segmentation. Finally, through the output mask image and the positioning algorithm the positioning of the picking points of tea buds and leaves is determined. One hundred tea tree bud and leaf pictures in a complex environment are selected for testing. The experimental results show that the average detection accuracy rate reaches 93.95% and that the recall rate reaches 92.48%. The tea picking location method presented in this paper is more versatile and robust in complex environments.

Abstract Although the classification method based on the deep neural network has achieved excellent results in classification tasks, it is difficult to apply to real-time scenarios because of high memory footprints and prohibitive inference times. Compared to unstructured pruning, structured pruning techniques can reduce the computation cost of the model runtime more effectively, but inevitably reduces the precision of the model. Traditional methods use fine tuning to restore model damage performance. However, there is still a large gap between the pruned model and the original one. In this paper, we use progressive multi-level distillation learning to compensate for the loss caused by pruning. Pre-pruning and post-pruning networks serve as the teacher and student networks. The proposed approach utilizes the complementary properties of structured pruning and knowledge distillation, which allows the pruned network to learn the intermediate and output representations of the teacher network, thus reducing the influence of the model subject to pruning. Experiments demonstrate that our approach performs better on CIFAR-10, CIFAR-100, and Tiny-ImageNet datasets with different pruning rates. For instance, GoogLeNet can achieve near lossless pruning on the CIFAR-10 dataset with 60% pruning. Moreover, this paper also proves that using the proposed distillation learning method during the pruning process achieves more significant performance gains than after completing the pruning.

In winter and spring, for greenhouses with larger areas and stereoscopic cultivation, distributed light environment regulation based on photosynthetic rate prediction model can better ensure good crop growth. In this paper, strawberries at flowering-fruit stage were used as the test crop, and the LI-6800 portable photosynthesis system was used to control the leaf chamber environment and obtain sample data by nested photosynthetic rate combination experiments under temperature, light and CO2 concentration conditions to study the photosynthetic rate prediction model construction method. For a small-sample, nonlinear real experimental data set validated by grey relational analysis, a photosynthetic rate prediction model was developed based on Support vector regression (SVR), and the particle swarm algorithm (PSO) was used to search the influence of the empirical values of parameters, such as the penalty parameter C, accuracy ε and kernel constant g, on the model prediction performance. The modeling and prediction results show that the PSO-SVR method outperforms the commonly used algorithms such as MLR, BP, SVR and RF in terms of prediction performance and generalization on a small sample data set. The research in this paper achieves accurate prediction of photosynthetic rate of strawberry and lays the foundation for subsequent distributed regulation of greenhouse strawberry light environment.

Abstract The landmark Net Zero Emissions by 2050 Scenario requires the revolution of today's energy system for realizing nonenergy‐related global economy. Advanced batteries with high energy density and safety are expected to realize the shift of end‐use sectors toward renewable and clean sources of electricity. Present Li‐ion technologies have dominated the modern energy market but face with looming challenges of limited theoretical specific capacity and high cost. Li–air(O2) battery, characterized by energy‐rich redox chemistry of Li stripping/plating and oxygen conversion, emerges as a promising “beyond Li‐ion” strategy. In view of the superior stability and inherent safety, a solid‐state Li–air battery is regarded as a more practical choice compared to the liquid‐state counterpart. However, there remain many challenges that retard the development of solid‐state Li–air batteries. In this review, we provide an in‐depth understanding of fundamental science from both thermodynamics and kinetics of solid‐state Li–air batteries and give a comprehensive assessment of the main challenges. The discussion of effective strategies along with authoritative demonstrations for achieving high‐performance solid‐state Li–air batteries is presented, including the improvement of cathode kinetics and durability, solid electrolyte design, Li anode optimization and protection, as well as interfacial engineering.

Strawberry maturity detection plays an essential role in modern strawberry yield estimation and robot-assisted picking and sorting. Due to the small size and complex growth environment of strawberries, there are still problems with existing recognition systems’ accuracy and maturity classifications. This article proposes a strawberry maturity recognition algorithm based on an improved YOLOv5s model named YOLOv5s-BiCE. This algorithm model is a replacement of the upsampling algorithm with a CARAFE module structure. It is an improvement on the previous model in terms of its content-aware processing; it also widens the field of vision and maintains a high level of efficiency, resulting in improved object detection capabilities. This article also introduces a double attention mechanism named Biformed for small-target detection, optimizing computing allocation, and enhancing content perception flexibility. Via multi-scale feature fusion, we utilized double attention mechanisms to reduce the number of redundant computations. Additionally, the Focal_EIOU optimization method was introduced to improve its accuracy and address issues related to uneven sample classification in the loss function. The YOLOv5s-BiCE algorithm was better at recognizing strawberry maturity compared to the original YOLOv5s model. It achieved a 2.8% increase in the mean average precision and a 7.4% increase in accuracy for the strawberry maturity dataset. The improved algorithm outperformed other networks, like YOLOv4-tiny, YOLOv4-lite-e, YOLOv4-lite-s, YOLOv7, and Fast RCNN, with recognition accuracy improvements of 3.3%, 4.7%, 4.2%, 1.5%, and 2.2%, respectively. In addition, we developed a corresponding detection app and combined the algorithm with DeepSort to apply it to patrol robots. It was found that the detection algorithm exhibits a fast real-time detection speed, can support intelligent estimations of strawberry yield, and can assist picking robots.

It is well known that crop classification is essential for genetic resources and phenotype development. Compared with traditional methods, convolutional neural networks can be utilized to identify features automatically. Nevertheless, crops and scenarios are quite complex, which makes it challenging to develop a universal classification method. Furthermore, manual design demands professional knowledge and is time-consuming and labor-intensive. In contrast, auto-search can create network architectures when faced with new species. Using rapeseed images for experiments, we collected eight types to build datasets (rapeseed dataset (RSDS)). In addition, we proposed a novel target-dependent search method based on VGGNet (target-dependent neural architecture search (TD-NAS)). The result shows that test accuracy does not differ significantly between small and large samples. Therefore, the influence of the dataset size on generalization is limited. Moreover, we used two additional open datasets (Pl@ntNet and ICL-Leaf) to test and prove the effectiveness of our method due to three notable features: (a) small sample sizes, (b) stable generalization, and (c) free of unpromising detections.

Surface-enhanced Raman spectroscopy (SERS) has attracted much attention because of its high sensitivity, high speed, and simple sample processing, and has great potential for application in the field of pesticide residue detection. However, SERS is susceptible to the influence of a complex detection environment in the detection of pesticide residues on the surface of fruits, facing problems such as interference from the spectral peaks of detected impurities, unclear dimension of effective correlation data, and poor linearity of sensing signals. In this work, the enhanced raw data of the pesticide thiram residues on the fruit surface using gold nanoparticle (Au-NPs) solution are formed into the raw data set of Raman signal in the IoT environment of Raman spectroscopy principal component detection. Considering the non-linear characteristics of sensing data, this work adopts kernel principal component analysis (KPCA) including radial basis function (RBF) to extract the main features for the spectra in the ranges of 653∼683 cm−1, 705∼728 cm−1, and 847∼872 cm−1, and discusses the effects of different kernel function widths (σ) to construct a qualitative analysis of pesticide residues based on SERS spectral data model, so that the SERS spectral data produce more useful dimensionality reduction with minimal loss, higher mean squared error for cross-validation in non-linear scenarios, and effectively weaken the interference features of detecting impurity spectral peaks, unclear dimensionality of effective correlation data, and poor linearity of sensing signals, reflecting better extraction effects than conventional principal component analysis (PCA) models.

Precise recognition of maize growth stages in the field is one of the critical steps in conducting precision irrigation and crop growth evaluation. However, due to the ever-changing environmental factors and maize growth characteristics, traditional recognition methods usually suffer from limitations in recognizing different growth stages. For the purpose of tackling these issues, this study proposed an improved U-net by first using a cascade convolution-based network as the encoder with a strategy for backbone network replacement to optimize feature extraction and reuse. Secondly, three attention mechanism modules have been introduced to upgrade the decoder part of the original U-net, which highlighted critical regions and extracted more discriminative features of maize. Subsequently, a dilation path of the improved U-net was constructed by integrating dilated convolution layers using a multi-scale feature fusion approach to preserve the detailed spatial information of in-field maize. Finally, the improved U-net has been applied to recognize different growth stages of maize in the field. The results clearly demonstrated the superior ability of the improved U-net to precisely segment and recognize maize growth stage from in-field images. Specifically, the semantic segmentation network achieved a mean intersection over union (mIoU) of 94.51% and a mean pixel accuracy (mPA) of 96.93% in recognizing the maize growth stage with only 39.08 MB of parameters. In conclusion, the good trade-offs made in terms of accuracy and parameter number demonstrated that this study could lay a good foundation for implementing accurate maize growth stage recognition and long-term automatic growth monitoring.

The “Dangshan” pear woolliness response is a physiological disease that mostly occurs in the pear growth process. The appearance of the disease is not obvious, and it is difficult to detect with the naked eye. Therefore, finding a way to quickly and nondestructively identify “Dangshan” pear woolliness disease is of great significance. In this paper, the near-infrared spectral (NIR) data of “Dangshan” pear samples were collected at 900–1700 nm reflectance spectra using a handheld miniature NIR spectrometer, and the data were modelled and analysed using random forest (RF), support vector machine (SVM) and boosting algorithms under the processing of 24 pretreatment methods. Considering the variations between different pretreatment methods, this work determined the relative optimality index of different pretreatment methods by evaluating their effects on model accuracy and Kappa and selected the best-performing first derivative with standard normal variate and Savitzky–Golay and first derivative with multiplicative scatter correction and Savitzky–Golay as the best pretreatment methods. With the best pretreatment method, all five models in the three categories showed good accuracy and stability after parameter debugging, with accuracy and F1 greater than 0.8 and Kappa floating at approximately 0.7, reflecting the good classification ability of the models and proving that near-infrared spectroscopy (NIRS) in the rapid identification of “Dangshan” pear woolliness response disease was feasible. By comparing the performance differences of the models before and after the pretreatment methods, it was found that the ensemble-learning models such as RF and boosting were more stringent on pretreatment methods in identifying “Dangshan” pear woolliness response disease than support vector machines, and the performance of the ensemble learning models was significantly improved under appropriate pretreatment methods. This experiment provided a relatively stable detection method for “Dangshan” pear woolliness response disease under nonideal detection conditions by analysing the impact of pretreatment methods and models on the prediction result.

Accurate diagnosis of pear tree nutrient deficiency symptoms is vital for the timely adoption of fertilization and treatment. This study proposes a novel method on the fused feature multi-head attention recording network with image depth and shallow feature fusion for diagnosing nutrient deficiency symptoms in pear leaves. First, the shallow features of nutrient-deficient pear leaf images are extracted using manual feature extraction methods, and the depth features are extracted by the deep network model. Second, the shallow features are fused with the depth features using serial fusion. In addition, the fused features are trained using three classification algorithms, F-Net, FC-Net, and FA-Net, proposed in this paper. Finally, we compare the performance of single feature-based and fusion feature-based identification algorithms in the nutrient-deficient pear leaf diagnostic task. The best classification performance is achieved by fusing the depth features output from the ConvNeXt-Base deep network model with shallow features using the proposed FA-Net network, which improved the average accuracy by 15.34 and 10.19 percentage points, respectively, compared with the original ConvNeXt-Base model and the shallow feature-based recognition model. The result can accurately recognize pear leaf deficiency images by providing a theoretical foundation for identifying plant nutrient-deficient leaves.

Vision-based underwater autonomous systems play a significant role in marine exploration. Stereo matching is one of the most popular applications for vision-based underwater autonomous systems, which recovers the geometric information of underwater scenes via stereo disparity estimation. While stereo matching in the air has achieved great progress with the development of neural networks, it generalizes poorly to the underwater scenario due to the challenging underwater degradation. In this paper, we propose a novel Multilevel Inverse Patchmatch Network (MIPNet) to iteratively model pair-wise correlations under underwater degradation and estimate stereo disparity with both local and global refinements. Specifically, we first utilized the inverse Patchmatch module in a novel multilevel pyramid structure to recover the detailed stereo disparity from the input stereo images. Secondly, we introduced a powerful Attentional Feature Fusion module to model pair-wise correlations with global context, ensuring high-quality stereo disparity estimation for both in-air and underwater scenarios. We evaluate the proposed method on the popular real-world ETH3D benchmark, and the highly competitive performance against the popular baselines demonstrates the effectiveness of the proposed method. Moreover, with its superior performance on our real-world underwater dataset, e.g., our method outperforms the popular baseline RAFT-Stereo by 27.1%, we show the good generalization ability of our method to underwater scenarios. We finally discuss the potential challenges for underwater stereo matching via our experiments on the impact of water.

Platinum nanoparticles (PtNPs) are classical peroxidase-like nanozyme; self-agglomeration of nanoparticles leads to the undesirable reduction in stability and catalytic activity. Herein, a hybrid peroxidase-like nanocatalyst consisting of PtNPs in situ growing on g–C3N4 nanosheets with enhanced peroxidase-mimic catalytic activity (PtNP@g–C3N4 nanosheets) was prepared for H2O2 and oxidase-based colorimetric assay. g–C3N4 nanosheets can be used as carriers to solve the problem of poor stability of PtNPs. We observed that the catalytic ability could be maintained for more than 90 days. PtNP@g–C3N4 nanosheets could quickly catalyze the oxidation of 3,3′,5,5′-tetramethylbenzidine (TMB), and the absorbance of blue color oxidized TMB (oxTMB) showed a robust linear relationship with the concentration of H2O2 (the detection limit (LOD): 3.33 μM). By utilizing H2O2 as a mediator, this strategy can be applied to oxidase-based biomolecules (glucose, organophosphorus, and so on, that generate or consume hydrogen peroxide) sensing. As a proof of concept, a sensitive assay of cholesterol that combined PtNP@g–C3N4 nanosheets with cholesterol oxidase (ChOx) cascade catalytic reaction was constructed with an LOD of 9.35 μM in a widespread range from 10 to 800 μM (R2 = 0.9981). In addition, we also verified its ability to detect cholesterol in fetal bovine serum. These results showed application prospect of PtNP@g–C3N4 nanosheets-based colorimetry in sensing and clinical medical detection.

The “Dangshan” pear woolliness response is a physiological disease that causes large losses for fruit farmers and nutrient inadequacies.The cause of this disease is predominantly a shortage of boron and calcium in the pear and water loss from the pear. This paper used the fusion of near-infrared Spectroscopy (NIRS) and Computer Vision Technology (CVS) to detect the woolliness response disease of “Dangshan” pears. This paper employs the merging of NIRS features and image features for the detection of “Dangshan” pear woolliness response disease. Near-infrared Spectroscopy (NIRS) reflects information on organic matter containing hydrogen groups and other components in various biochemical structures in the sample under test, and Computer Vision Technology (CVS) captures image information on the disease. This study compares the results of different fusion models. Compared with other strategies, the fusion model combining spectral features and image features had better performance. These fusion models have better model effects than single-feature models, and the effects of these models may vary according to different image depth features selected for fusion modeling. Therefore, the model results of fusion modeling using different image depth features are further compared. The results show that the deeper the depth model in this study, the better the fusion modeling effect of the extracted image features and spectral features. The combination of the MLP classification model and the Xception convolutional neural classification network fused with the NIR spectral features and image features extracted, respectively, was the best combination, with accuracy (0.972), precision (0.974), recall (0.972), and F1 (0.972) of this model being the highest compared to the other models. This article illustrates that the accuracy of the “Dangshan” pear woolliness response disease may be considerably enhanced using the fusion of near-infrared spectra and image-based neural network features. It also provides a theoretical basis for the nondestructive detection of several techniques of spectra and pictures.

Pine wilt disease (PWD) is a great danger, due to two aspects: no effective cure and fast dissemination. One key to the prevention and treatment of pine wilt disease is the early detection of infected wood. Subsequently, appropriate treatment can be applied to limit the further spread of pine wilt disease. In this work, a UAV (Unmanned Aerial Vehicle) with a RGB (Red, Green, Blue) camera was employed as it provided high-quality images of pine trees in a timely manner. Seven flights were performed above seven sample plots in northwestern Beijing, China. Then, raw images captured by the UAV were further pre-processed, classified, annotated, and formed the research datasets. In the formal analysis, improved YOLOv5 frameworks that integrated four attention mechanism modules, i.e., SE (Squeeze-and-Excitation), CA (Coordinate Attention), ECA (Efficient Channel Attention), and CBAM (Convolutional Block Attention Module), were developed. Each of them had been shown to improve the overall identification rate of infected trees at different ranges. The CA module was found to have the best performance, with an accuracy of 92.6%, a 3.3% improvement over the original YOLOv5s model. Meanwhile, the recognition speed was improved by 20 frames/second compared to the original YOLOv5s model. The comprehensive performance could well support the need for rapid detection of pine wilt disease. The overall framework proposed by this work shows a fast response to the spread of PWD. In addition, it requires a small amount of financial resources, which determines the duplication of this method for forestry operators.

In this paper, a novel image splicing detection and localization scheme is proposed based on the local feature descriptor which is learned by deep convolutional neural network (CNN). A two-branch CNN, which serves as an expressive local descriptor is presented and applied to automatically learn hierarchical representations from the input RGB color or grayscale test images. The first layer of the proposed CNN model is used to suppress the effects of image contents and extract the diverse and expressive residual features, which is deliberately designed for image splicing detection applications. In specific, the kernels of the first convolutional layer are initialized with an optimized combination of the 30 linear high-pass filters used in calculation of residual maps in spatial rich model (SRM), and is fine-tuned through a constrained learning strategy to retain the high-pass filtering properties for the learned kernels. Both the contrastive loss and cross entropy loss are utilized to jointly improve the generalization ability of the proposed CNN model. With the block-wise dense features for a test image extracted by the pre-trained CNN-based local descriptor, an effective feature fusion strategy, known as block pooling, is adopted to obtain the final discriminative features for image splicing detection with SVM. Based on the pre-trained CNN model, an image splicing localization scheme is further developed by incorporating the fully connected conditional random field (CRF). Extensive experimental results on several public datasets show that the proposed CNN based scheme outperforms some state-of-the-art methods not only in image splicing detection and localization performance, but also in robustness against JPEG compression.

Precise pear detection and recognition is an essential step toward modernizing orchard management. However, due to the ubiquitous occlusion in orchards and various locations of image acquisition, the pears in the acquired images may be quite small and occluded, causing high false detection and object loss rate. In this paper, a multi-scale collaborative perception network YOLOv5s-FP (Fusion and Perception) was proposed for pear detection, which coupled local and global features. Specifically, a pear dataset with a high proportion of small and occluded pears was proposed, comprising 3680 images acquired with cameras mounted on a ground tripod and a UAV platform. The cross-stage partial (CSP) module was optimized to extract global features through a transformer encoder, which was then fused with local features by an attentional feature fusion mechanism. Subsequently, a modified path aggregation network oriented to collaboration perception of multi-scale features was proposed by incorporating a transformer encoder, the optimized CSP, and new skip connections. The quantitative results of utilizing the YOLOv5s-FP for pear detection were compared with other typical object detection networks of the YOLO series, recording the highest average precision of 96.12% with less detection time and computational cost. In qualitative experiments, the proposed network achieved superior visual performance with stronger robustness to the changes in occlusion and illumination conditions, particularly providing the ability to detect pears with different sizes in highly dense, overlapping environments and non-normal illumination areas. Therefore, the proposed YOLOv5s-FP network was practicable for detecting in-field pears in a real-time and accurate way, which could be an advantageous component of the technology for monitoring pear growth status and implementing automated harvesting in unmanned orchards.

Due to the explosive growth of image data, image annotation has been one of the most popular research directions in computer vision. It has been widely used in image retrieval, image analysis and understanding. Because traditional manual image annotation is time consuming, more advanced automatic annotation methods are needed. A major challenge in developing an efficient image annotation method is how to effectively use all available information contained in the data. To this end, this paper proposes a novel image annotation framework that uses multiple information from data. It employs nonnegative matrix tri-factorization (NMTF) to simultaneously factorize image-to-label, image-to-feature, and feature-to-label relation matrices using their intertype relationships and incorporates the intratype information through manifold regularizations. This method can be referred to as the trigraph regularized collective matrix tri-factorization framework (TG-CMTF). TG-CMTF captures the correlations among different labels, different images and different features. By taking advantage of these relations from images, features and labels, TG-CMTF can achieve better annotation performance than most state-of-the-art methods. The promising experimental results on three standard benchmarks have shown the effectiveness of this information. Furthermore, we show the annotation process as a precise optimization problem and solve it by an iterative algorithm, which proves the correctness of the proposed method from the mathematical theory.

Glycyrrhetinic acid (GA) is an anti-inflammatory drug with potential for development. However, the poor solubility of GA in water leads to extremely low bioavailability, which limits its clinical applications. Solid dispersions have become some of the most effective strategies for improving the solubility of poorly soluble drugs. Soluplus®, a non-cytotoxic amphiphilic solubilizer, significantly improves the solubility of BCS II drugs and improves the bioavailability of insoluble drugs. l-arginine (L-Arg) can be used as a small molecular weight excipient to assist in improving the solubility of insoluble drugs. In this study, we developed a new formulation for oral administration by reacting GA and L-Arg to form salts by co-solvent evaporation and then adding the polymer-solvent Soluplus® with an amphiphilic chemical structure to prepare a solid dispersion GA-SD. The chemical and physical properties of GA-SD were characterized by DLS, TEM, XRD, FT-IR and TG. The anti-inflammatory activity of GA-SD was verified by LPS stimulation of RAW 267.5 cells simulating a cellular inflammation model, TPA-induced ear edema model in mice, and ethanol-induced gastric ulcer model. The results showed that the amide bond and salt formation of GA-SD greatly improved GA solubility. GA-SD effectively improved the anti-inflammatory effect of free GA in vivo and in vitro, and GA-SD had no significant effect on liver and kidney function, no significant tissue toxicity, and good biosafety. In conclusion, GA-SD with L-Arg and Soluplus® is an effective method to improve the solubility and bioavailability of GA. As a safe and effective solid dispersion, it is a promising anti-inflammatory oral formulation and provides some references for other oral drug candidates with low bioavailability.

The number and growth of new shoots are very important information for bamboo forest cultivation and management. At present, there is no real-time, efficient and accurate monitoring method. In this study, a fixed webcam was applied for image capture, optimized YOLOv4 was used to model the detection of moso bamboo shoots, and a strategy of sorting and screening was proposed to track each moso bamboo shoot. The change in the number and height of moso bamboo shoots was obtained according to the number and height of detection boxes. The experimental results show that the system can remotely and automatically obtain the number of moso bamboo shoots and the pixel height of each bamboo shoot at any given time. The average relative error and variance in the number of moso bamboo shoots were 1.28% and 0.016%, respectively, and those for the corresponding pixel height results were −0.39% and 0.02%. This system can be applied to a series of monitoring purposes, such as the daily or weekly growth rate of moso bamboo shoots at monitoring stations and trends in the height of selected bamboo shoots.

Due to the complex environments in real fields, it is challenging to conduct identification modeling and diagnosis of plant leaf diseases by directly utilizing in-situ images from the system of agricultural Internet of things. To overcome this shortcoming, one approach, based on small sample size and deep convolutional neural network, was proposed for conducting the recognition of cucumber leaf diseases under field conditions. One two-stage segmentation method was presented to acquire the lesion images by extracting the disease spots from cucumber leaves. Subsequently, after implementing rotation and translation, the lesion images were fed into the activation reconstruction generative adversarial networks for data augmentation to generate new training samples. Finally, to improve the identification accuracy of cucumber leaf diseases, we proposed dilated and inception convolutional neural network that was trained using the generated training samples. Experimental results showed that the proposed approach achieved the average identification accuracy of 96.11% and 90.67% when implemented on the data sets of lesion and raw field diseased leaf images with three different diseases of anthracnose, downy mildew, and powdery mildew, significantly outperforming those existing counterparts, indicating that it offered good potential of serving field application of agricultural Internet of things.

Tight oil reservoirs are characterized by a low porosity, low permeability, and strong heterogeneity. The macropores, throats, and microcracks in reservoirs are the main seepage channels, which affect the seepage law in the reservoirs. In particular, oil-water two-phase flow in different types of pores requires further study. In this study, two groups of online NMR displacement experiments were designed to study the seepage characteristics of tight oil reservoirs. It was found that the main seepage channels for oil-water two-phase flow are the microcracks, large pores, and throats in the reservoir. The large pores are mainly micron and submicron scale in size. The oil in the small pores is only transferred to the large pores through imbibition to participate in the flow, and there is no two-phase flow. Based on the influence of different pore structures on the seepage law of a tight reservoir, the pores were divided into seepage zones, and a multistage seepage model for tight reservoirs was established. Based on this model, the effects of the imbibition, stress sensitivity, threshold pressure gradient, and Jamin effect on model’s yield were studied. The results show that imbibition is no longer effective after a while. Owing to the stress sensitivity, the threshold pressure gradient, and the Jamin effect, oil production will be reduced. As the parameter value increases, the oil production decreases. The production decreases rapidly in the early stage of mining while decreases slowly in the later stage, exhibiting a trend of high yield in the early stage and stable yield in the later stage.

Intensive animal husbandry is becoming more and more popular with the adoption of modern livestock farming technologies. In such circumstances, it is required that the welfare of animals be continuously monitored in a real-time way. To this end, this study describes one on-farm welfare monitoring system for goats, with a combination of Internet of Things and machine learning. First, the system was designed for uninterruptedly monitoring goat growth in a multifaceted and multilevel manner, by means of collecting on-farm videos and representative environmental data. Second, the monitoring hardware and software systems were presented in detail, aiming at supporting remote operation and maintenance, and convenience for further development. Third, several key approaches were put forward, including goat behavior analysis, anomaly data detection, and processing based on machine learning. Through practical deployment in the real situation, it was demonstrated that the developed system performed well and had good potential for offering real-time monitoring service for goats’ welfare, with the help of accurate environmental data and analysis of goat behavior.

Deep neural networks (DNNs) have become the de facto standard for image recognition tasks, and their applications with respect to plant diseases have also obtained remarkable results. However, the large number of parameters and high computational complexities of these network models make them difficult to deploy on farms in remote areas. In this paper, focusing on the problems of resource constraints and plant diseases, we propose a DNN-based compression method. In order to reduce computational burden, this method uses lightweight fully connected layers to accelerate reasoning, pruning to remove redundant parameters and reduce multiply–accumulate operations, knowledge distillation instead of retraining to restore the lost accuracy, and then quantization to compress the size of the model further. After compressing the mainstream VGGNet and AlexNet models, the compressed versions are applied to the Plant Village dataset of plant disease images, and a performance comparison of the models before and after compression is obtained to verify the proposed method. The results show that the model can be compressed to 0.04 Mb with an accuracy of 97.09%. This experiment also proves the effectiveness of knowledge distillation during the pruning process, and compressed models are more efficient than prevalent lightweight models.

It is necessary to carry out experiments on cores with different degrees of crack development when studying the seepage law of cracked reservoirs and evaluating cracks. The seepage experiment in the laboratory requires cores with different degrees of microcrack development; cores obtained via conventional drilling cannot meet the requirements, and the efficacies and evaluation methods of geological parameters used for artificial cracks are not perfect. In this study, cores are loaded using a triaxial gripper, and cracks are produced by changing the difference of stress; the relationship between the increased rate of permeability and the change in stress concentration is used to evaluate the degree of development of the crack in real time. The angle between the cracks and the maximum principal stress direction, calculated using the Mohr–Coulomb failure criterion, is 20–27.5°, which provides theoretical support for the process of crack creation. The experimental results show that the permeability variation curve shows two obvious turning points, which divide the whole zone into a reduction zone, a slow increase zone, and a rapid increase zone. Through the obtained experimental and evaluation results, a complete system for crack creation and evaluation is established, which can provide strong support for the study of cracked reservoirs.

The separation of solution gas has great influence on the development of gas-bearing tight oil reservoirs. In this study, physical simulation and high-pressure mercury intrusion were used to establish a method for determining the porous flow resistance gradient of gas-bearing tight oil reservoirs. A mathematical model suitable for injection–production well networks is established based on the streamline integral method. The concept of pseudo-bubble point pressure is proposed. The experimental results show that as the back pressure decreases from above the bubble point pressure to below the bubble point pressure, the solution gas separates out. During this process, the porous flow resistance gradient is initially equal to the threshold pressure gradient of the oil single-phase fluid, then it becomes relatively small and stable, and finally it increases rapidly and exponentially. The lower the permeability, the higher the pseudo-bubble point pressure, and the higher the resistance gradient under the same back pressure. For tight reservoirs, the production pressure should be maintained above the pseudo-bubble point pressure when the permeability is lower than a certain value. When the permeability is higher than a certain value, the pressure can be reduced below the pseudo-bubble point pressure, and there is a reasonable range. The mathematical results show that after degassing, the oil production rate and the effective utilization coefficient of oil wells decline rapidly. These declines occur later and have a flat trend for high permeability formations, and the production well pressure can be reduced to a lower level. Fracturing can effectively increase the oil production rate after degassing. A formation that cannot be utilized before fracturing because of the blocked throats due to the separation of the solution gas can also be utilized after fracturing. When the production well pressure is lower than the bubble point pressure, which is not too large, the fracturing effect is better.

Recent advances in agriculture-oriented sensors and wireless communication technologies offer great opportunities for the development and application of precision agriculture. This article describes our experience during the deployment of an experimental in-field sensor network that provided real-time monitoring in a 10-year-old date palm orchard. Taking into account the spatial variability of multiple growth-determining soil variables, the sub-regions of the sensor placement were determined by delineating the orchard into several monitoring zones. Then, the specific number of the required sensors was calculated based on the requirement of data accuracy. The sensors were distributed into the target orchard by combining the tree number with the coefficient variability of growth-determining soil variables in the monitoring zones. Finally, the networking was realized by means of synthesizing the received wireless signal strength distribution, the layout of monitoring zones, and the number of the required sensors. Some data from the deployed wireless sensor network are presented and discussed.

Some frame components, such as SYNC (frame synchronization) and RTS/CTS (Ready to Send/Clear to Send), are not taken into consideration when the traditional setting strategies conduct the optimization of SMAC (Sensor MAC) contention window size. This paper proposes mathematical models that allow the analysis of data packets forwarding delay within one SMAC virtual cluster. Simulation results in OMNeT++ show good agreements with the proposed mathematical models, validating the models' correctness. The curve analyses of the models confirm the existence of delay-optimization-oriented contention window size that is closely related to the number of simultaneously contending nodes. Afterwards, it is shown that SYNC, RTS/CTS and EIFS (Extended InterFrame Space) have impacts on the optimal contention window size and expected delivery delay to various degrees, as well as throughput and energy efficiency. One ideal setting strategy of delay-optimization-oriented SMAC contention window size requires the combination of the network scale, SYNC, RTS/CTS and EIFS. Additionally, it is demonstrated that the proposed setting strategy makes contributions to the improvement in the existing SMAC extensions when they are integrated with each other, in terms of the end-to-end delay, throughput and energy consumption.

Due to the limited energy budget, great efforts have been made to improve energy efficiency for wireless sensor networks. The advantage of compressed sensing is that it saves energy because of its sparse sampling; however, it suffers inherent shortcomings in relation to timely data acquisition. In contrast, prediction-based approaches are able to offer timely data acquisition, but the overhead of frequent model synchronization and data sampling weakens the gain in the data reduction. The integration of compressed sensing and prediction-based approaches is one promising data acquisition scheme for the suppression of data transmission, as well as timely collection of critical data, but it is challenging to adaptively and effectively conduct appropriate switching between the two aforementioned data gathering modes. Taking into account the characteristics of data gathering modes and monitored data, this research focuses on several key issues, such as integration framework, adaptive deviation tolerance, and adaptive switching mechanism of data gathering modes. In particular, the adaptive deviation tolerance is proposed for improving the flexibility of data acquisition scheme. The adaptive switching mechanism aims at overcoming the drawbacks in the traditional method that fails to effectively react to the phenomena change unless the sampling frequency is sufficiently high. Through experiments, it is demonstrated that the proposed scheme has good flexibility and scalability, and is capable of simultaneously achieving good energy efficiency and high-quality sensing of critical events.

In this study, we address the problem of selecting the optimal end-to-end paths for link loss inference in order to improve the performance of network tomography applications, which infer the link loss rates from the path loss rates. Measuring the path loss rates using end-to-end probing packets may incur additional traffic overheads for networks, so it is important to select the minimum path set carefully while maximizing their performance. The usual approach is to select the maximum independent paths from the candidates simultaneously, while the other paths can be replaced by linear combinations of them. However, this approach ignores the fact that many paths always exist that do not lose any packets, and thus it is easy to determine that all of the links of these paths also have 0 loss rates. Not considering these good paths will inevitably lead to inefficiency and high probing costs. Thus, we propose an adaptive path selection method that selects paths sequentially based on the loss rates of previously selected paths. We also propose a theorem as well as a graph construction and decomposition approach to efficiently find the most valuable path during each round of selection. Our new method significantly outperforms the classical path selection method based on simulations in terms of the probing cost, number of accurate links determined, and the running speed.