Search

The kale crop is an important bulk vegetable, and automatic segmentation to recognize kale is fundamental for effective field management. However, complex backgrounds and texture-rich edge details make fine segmentation of kale difficult. To this end, we constructed a kale dataset in a real field scenario and proposed an UperNet semantic segmentation model with a Swin transformer as the backbone network and improved the model according to the growth characteristics of kale. Firstly, a channel attention module (CAM) is introduced into the Swin transformer module to improve the representation ability of the network and enhance the extraction of kale outer leaf and leaf bulb information; secondly, the extraction accuracy of kale target edges is improved in the decoding part by designing an attention refinement module (ARM); lastly, the uneven distribution of classes is solved by modifying the optimizer and loss function to solve the class distribution problem. The experimental results show that the improved model in this paper has excellent performance in feature extraction, and the average intersection and merger ratio (mIOU) of the improved kale segmentation can be up to 91.2%, and the average pixel accuracy (mPA) can be up to 95.2%, which is 2.1 percentage points and 4.7 percentage points higher than the original UperNet model, respectively, and it effectively improves the segmentation recognition of kale. [ABSTRACT FROM AUTHOR]

In recent years, the increasing demand for knowledge services and the challenges of information overload have posed significant problems in delivering personalized and efficient agricultural knowledge services. This paper presents a comprehensive framework that addresses the issues of vague user positioning, serious privacy leakage, and low efficiency in personalized knowledge services within the national agricultural knowledge intelligent service cloud platform. The proposed framework utilizes privacy-protected user portraits based on generative adversarial nets (GAN) and leverages the TextCNN-LSTM algorithm for agricultural knowledge service prediction. By embedding labels into the algorithm and employing data obfuscation techniques, the framework achieves accurate inference of user behavior while preserving user privacy. Experimental results demonstrate the effectiveness and accuracy of the proposed framework, highlighting its potential for regional precise positioning and recommendation of personalized agricultural knowledge services. Experimental data shows that the average absolute error and root-mean-square error of this method are 1.1997 and 1.4143, respectively, and compared with MLP, TextCNN, and LSTM models, and it has higher prediction accuracy. In recent years, the increasing demand for knowledge services and the challenges of information overload have posed significant problems in delivering personalized and efficient agricultural knowledge services. [ABSTRACT FROM AUTHOR]

Wireless sensor networks (WSN) enable the acquisition of multisource environmental data and crop states in precision agriculture. However, the complex agricultural environment causes the WSN topology to change frequently and link connection probability is difficult to predict. In order to improve the utilization of network resources and balance the network energy consumption, this paper studies an agricultural fault-tolerant topology construction method based on the potential game and cut vertex detection. Considering the connectivity redundancy, node lifetime, and residual energy, a fault-tolerant topology algorithm for agricultural WSN based on a double price function is designed. The network is clustered according to the node location and residual energy to form a single-hop effective cluster. Based on the network cluster, the price function is constructed in order to reduce energy consumption and balance network energy efficiency. The initial transmit power set supporting inter-cluster communication is obtained by potential game theory. While preserving the game characteristics of topology, the redundant links are eliminated and the transmit power is adjusted by a cut vertex detection algorithm to realize the construction of a 2-connected cluster head network. Simulation results show that the network topology constructed by the studied algorithm can balance the energy consumption and prolong the network lifetime effectively. [ABSTRACT FROM AUTHOR]

Developing small-angle neutron scattering techniques at compact accelerator-driven neutron sources (CANS) is of great importance for expanding the user community and advancing CANS capability. At the Compact Pulsed Hadron Source (CPHS) at Tsinghua University, neutron-focusing mirrors are under intensive research to address the challenge. A grazing-incidence focusing SANS (gif-SANS) project is initialized. It employs a nested supermirror assembly with a large collecting area to achieve ⩾ 10 5 n/s neutron intensity at Q min ⩽ 0.007 Å − 1 . It will equip two detectors, one being a 3He detector for normal Q-range measurements, and the other being a high-resolution detector for extending the Q min down to 10 − 3 Å − 1 . In this work, we present the conceptual design of the gif-SANS at CPHS. Such a scheme is conducive to enable high-performance SANS measurements at CANS. [ABSTRACT FROM AUTHOR]

Wireless sensor network (WSN) is the key sensing resource for the internet of things (IoT) in vegetable greenhouse. The coverage control ensures that WSN can obtain enough effective information. However, the current coverage researches ignore the object size and lack of attention to the occlusion between targets. There are many leaves and fruits in vegetables, which can easily cause blind area and low utilization of directional sensors. Based on the geometric relationship between the directional sensors and targets, this paper studies a non-occlusion coverage scheme for the greenhouse IoT. Firstly, combined with the traditional coverage theory, a directional coverage model without occlusion is constructed by analysing the multivariate relationship between the sensor nodes and monitored targets. An objective function is then established to maximize the effective coverage. Based on the directional coverage model, this paper studies a hierarchical cooperative particle swarm optimization algorithm, which decomposes the global effective coverage problem into the utilization optimization of each sensor and finally get the orientation angle set. The experimental results show that the studied model and algorithm can avoid occlusion between covered objects while improving sensor utilization to a certain degree. [ABSTRACT FROM AUTHOR]

Neutron current is crucial for neutron scattering measurements, especially for those at compact sources. Among various neutron optical devices, neutron-focusing mirrors have been drawing increasing attention due to their great potential in enhancing neutron current on sample for small-angle neutron scattering instruments. The state-of-the-art neutron-focusing mirrors, which can collect a large divergence of neutron beam, are required to be nestable and simultaneously be able to be coated with supermirrors. In this paper, we employ the nested conical optics to manufacture nested neutron-focusing mirrors with supermirror coatings. This technique enables such mirrors by assembling partial cylindrical glass segments into nested shells. A prototype system, containing two shell conical mirrors with m =2 Ni/Ti supermirror coatings, has been designed and manufactured. Ray-tracing simulations and neutron beam experiments have been carried out to evaluate its focusing performance and neutron reflectivity. The results show that neutron current collected by this prototype system is about 26 times higher than that by a pinhole design with the same Q min. Moreover, we discuss some possible further improvements on the fabrication process. [ABSTRACT FROM AUTHOR]

The neutron flux of the Compact Pulsed Hadron Source (CPHS) is about 2–3 orders of magnitude lower than that of large neutron sources, which means that the beam intensity should be improved to achieve good statistics. Multi‐pinhole collimation can be used to obtain a lower Q with an acceptable beam intensity in a very small angle neutron scattering (VSANS) instrument and a higher beam intensity for a larger sample size in a small‐angle neutron scattering (SANS) instrument. A new nine‐pinhole structure is used in a SANS instrument at CPHS to achieve an acceptable range and resolution of Q and a higher beam intensity compared to single‐pinhole collimation. The crosstalk issue associated with multi‐pinhole collimation is addressed using an optimized algorithm to achieve a higher safety margin and a larger pinhole size with a higher beam intensity at the sample. Different collimator aperture structures are compared on the basis of their noise production. Experiments are performed to verify the theory of calculating reflection noise from the inner surface of the collimator's aperture and parasitic noise from the beveled collimator structure. From a simulated SANS experiment using cold neutrons in the SANS instrument, it is clarified that multi‐pinhole collimators with an opening angle on the downstream side have better performance than those with an opening angle on the upstream side and straight‐cut collimators. Compared with a single‐pinhole collimation system, a nine‐pinhole collimation system increases the intensity at the sample by approximately sevenfold when the sample size is increased by 20‐fold for CPHS‐SANS, and the signal‐to‐noise ratio is improved by exploiting a specific collimator aperture structure. Our goal is to install a multi‐pinhole collimator based SANS instrument at CPHS in the future, and it is hoped that these results will serve to promote the utilization of multi‐pinhole collimation systems at other facilities. A small‐angle neutron scattering instrument based on a multi‐pinhole collimator system is used to increase neutron current at the sample compared with a single‐pinhole collimator system. Disadvantages such as crosstalk and reflection noise are overcome by the described optimization. [ABSTRACT FROM AUTHOR]

Density functional theory calculations are carried out to investigate the adsorption characteristics of methane (CH 4 ), carbon dioxide (CO 2 ), hydrogen (H 2 ), hydrogen sulfide (H 2 S), nitrogen (N 2 ), and oxygen (O 2 ) on the surface of pyridine-like nitrogen doped graphene (PNG) as well as noble metal (Rh, Pt, Pd) decorated PNG to elaborate their potentials as gas sensors. The adsorption intensities of biogas on noble metal (Rh, Pt, Pd) decorated PNG are in the order of O 2  > H 2 S   > N 2  > CH 4  > CO 2  > H 2 , which are corresponded to the order of their sensitivity on surface. Compared with biogas adsorption on pristine PNG, there exist higher adsorption ability, higher charge transfer and higher orbital hybridization upon adsorption on noble metal (Rh, Pt, Pd) decorated PNG. Consequently, the noble metal (Rh, Pt, Pd) decorated PNG can transform the existence of CH 4 , CO 2 , H 2 , H 2 S, N 2 , and O 2 molecules into electrical signal and they could potentially be used as ideal sensors for detection of biogas in ambient situation. [ABSTRACT FROM AUTHOR]

Density functional theory calculations are carried out to study the adsorption of mercury and arsenic on Pd n (n = 1–6) supported on pyridine-like nitrogen doped graphene (PNG). Owing to the promising sensitivity in trace amounts of atoms or molecules, PNG can be acted as micro-sensor for sensing heavy metals in agriculture soils. Through the analyses of structural and electronic properties of pristine PNG and Pd atom decorated PNG, we find that the most favorable adsorption site for Pd atom is the vacancy site. The analyses of structural and electronic properties reveal that the Pd atom or clusters can enhance the reactivity for Hg and AsH 3 adsorption on PNG. The adsorption ability of Hg on Pd n decorated PNG is found to be related to the d-band center (ε d ) of the Pd n , in which the closer ε d of Pd n to the Fermi level, the higher adsorption strength for Hg on Pd n decorated PNG. Moreover, the charge transfer between Pd n and arsenic may constitute arsenic adsorption on Pd n decorated PNG. Further design of highly efficient carbon based sorbents for heavy metals removal should be focused on tailoring ε d of adsorbed metals. [ABSTRACT FROM AUTHOR]

Based on density functional theory calculations, we investigate the boron substituted graphyne in terms of Li storage as well as metal-decorated graphyne and graphyne-like boron nitride layer (BN-yne) for hydrogen storage. Compared with pristine graphyne, the electric storage capacity and lithiation potentials of B-doped graphyne are improved significantly. The result can provide helpful information for the design and fabrication of anode materials of Lithium-ion battery, which is of crucial for the lifetime of wireless sensor nodes. Through the analyses of structural and electronic properties of graphyne and its boron nitride analog, we find that the most preferable adsorptive site of M (Ni, Cr, Pd) atom is the hollow site of 12-atom hexagonal ring. For hydrogen adsorption on M (Ni, Cr, Pd)-decorated graphyne and BN-yne, the hydrogen are not dissociated and all stored in molecular form. Due to the steric hindrance of the H 2 molecules, each M (Ni, Cr, Pd) atom can adsorb up to three H 2 molecules. The analyses of electronic properties show that the interactions between M (Ni, Cr, Pd) atom and H 2 molecule are the hybridization between them and the polarization induced by electrostatic field of M (Ni, Cr, Pd) atoms. Our work demonstrates that metal decorated graphyne and its boron nitride analog are promising for energy storage. [ABSTRACT FROM AUTHOR]

Y3Fe5O12 ceramics are prepared by the solid state reaction method. Grain size and Fe2+ concentration-dependent magnetic, dielectric, and magnetodielectric properties are investigated systematically. It is found that grain size and Fe2+ concentration increase with increasing sintering temperature, which thus results in increased saturation magnetization while decreased coercive field and remanent magnetization. Three sets of dielectric relaxation are observed in temperature range from 10 to 500 K, which are considered to be originated from quantum tunneling effect of electrons, electron hopping between Fe2+ and Fe3+, and the migration of oxygen vacancies, respectively. Cole-Cole fitting results of relaxation II indicate that a correlation exists among the hopping electrons, and the correlation is strengthened as Fe2+ concentration increases, which in turn results in easier hopping process and decreased activation energy. More interesting, YIG ceramics sintered at 1350 °C show both considerable magnetic and dielectric properties at room temperature, opening a new window for the application of YIG as a possible multi-susceptibile single-phase material. [ABSTRACT FROM AUTHOR]

Target objects are direct characteristics of image description, but there are significant limitations, including the image being sensitive to changes in global illumination, etc., which lead to lack direct descriptive ability for objects in an image. Image target segmentation is one of the primary image processing operations, which means it is the first operation of the processing of an image. For the segmentation of a target area from the image, there are two kinds of existing methods. One method is to manually remove the obstructions to be a uniform color fill or outline its contours; and another way is to use the shortest path algorithm or the expansion of biology technique to extract optimal edges. For an image with unclear target boundaries and background, it is very necessary to eliminate the interfering information by transforming the interference information into fuzzy sets. In this paper, to solve the complex background images and the image target segmentation error, considering the dominance on image processing of fuzzy sets, the maximum fuzzy correlation criterion is proposed. Image is converted to a fuzzy domain, and then the maximum fuzzy correlation criterion is calculated. It is important to optimize the image maximum fuzzy membership function. Finally, experiment on leaf and fruit agricultural images, the results show that more edge details can be detected by fuzzy sets. [ABSTRACT FROM AUTHOR]

Three-dimensional (3D) printing technology is advantageous in the fast prototyping of complex structures, but its utilization in functional material fabrication is still limited due to a lack of activation capability. To fabricate and activate the functional material of electrets, a synchronized 3D printing and corona charging method is presented to prototype and polarize polylactic acid electrets in one step. By upgrading the 3D printer nozzle and incorporating a needle electrode to apply high voltage, parameters such as needle tip distance and applied voltage level were compared and optimized. Under different experimental conditions, the average surface distribution in the center of the samples was −1498.87 V, −1115.73 V, and −814.51 V. Scanning electron microscopy results showed that the electric field contributes to keeping the printed fiber structure straight. The polylactic acid electrets exhibited relatively uniform surface potential distribution on a sufficiently large sample surface. In addition, the average surface potential retention rate was improved by 12.021-fold compared to ordinary corona-charged samples. The above advantages are unique to the 3D-printed and polarized polylactic acid electrets, proving that the proposed method is suitable for quickly prototyping and effectively polarizing the polylactic acid electrets simultaneously. [ABSTRACT FROM AUTHOR]

(Dicyclopentadiene) platinum(II) dichloride was found to be an efficient hydrosilylation catalyst (homogeneous) upon a wide variety of functionalized alkenes and alkenes terminated with chemical moieties (diphenyl amino-, N -carbazol- and N -isoindoline-1,3-dione-). It is noteworthy that the hydrosilylation of aminated alkenes with triethoxysilane exhibited the yield of over 70% and the selectivity (γ-isomer/β-isomer) of more than 3/1. Due to steric hindrance lowering Markovnikov probability, the alkenes with big terminal moieties (diphenyl amino-, N -carbazol- and N -isoindoline-1,3-dione-) presented the high ratio of anti-Markovnikov isomers. The strategy of the hydrosilylation of the protected diamino chelating alkene was developed. [ABSTRACT FROM AUTHOR]

The magnetocaloric (MC) properties in various magnetic materials were well determined within last three decades due to its own potential application in the field of the solid-state magnetic refrigeration (MR) technology as well as for a better understanding of the physical properties of the materials. In this work, a single phased ternary compound, the Gd2NiGa3, has been synthesized by conventional arc-melting process and studied in detail with regard to its magnetic and cryogenic MC properties. The Gd2NiGa3 compound undergoes a first-ordered magnetic transition (MT) around ~ 26 K from antiferromagnetic to paramagnetic (AFM to PM) state. Considerable cryogenic reversible MC effect has been found in the Gd2NiGa3 compound, the MC parameters mainly include the maximum magnetic isothermal entropy changes (− ΔSMmax), and the refrigerant capacity (RC) (ΔH of 0–50 kOe) are 6.82 J/kgK and 256.2 J/kg, respectively. These MC parameters are considerably large when compared to those of the reported MC materials, indicating that the Gd2NiGa3 compound also has potential applications for the cryogenic MR technology. [ABSTRACT FROM AUTHOR]

The governance of rural living environments is one of the important tasks in the implementation of a rural revitalization strategy. At present, the illegal behaviors of random construction and random storage in public spaces have seriously affected the effectiveness of the governance of rural living environments. The current supervision on such problems mainly relies on manual inspection. Due to the large number and wide distribution of rural areas to be inspected, this method is limited by obvious disadvantages, such as low detection efficiency, long-time spending, and huge consumption of human resources, so it is difficult to meet the requirements of efficient and accurate inspection. In response to the difficulties encountered, a low-altitude remote sensing inspection method on rural living environments was proposed based on a modified YOLOv5s-ViT (YOLOv5s-Vision Transformer) in this paper. First, the BottleNeck structure was modified to enhance the multi-scale feature capture capability of the model. Then, the SimAM attention mechanism module was embedded to intensify the model's attention to key features without increasing the number of parameters. Finally, the Vision Transformer component was incorporated to improve the model's ability to perceive global features in the image. The testing results of the established model showed that, compared with the original YOLOv5 network, the Precision, Recall, and mAP of the modified YOLOv5s-ViT model improved by 2.2%, 11.5%, and 6.5%, respectively; the total number of parameters was reduced by 68.4%; and the computation volume was reduced by 83.3%. Relative to other mainstream detection models, YOLOv5s-ViT achieved a good balance between detection performance and model complexity. This study provides new ideas for improving the digital capability of the governance of rural living environments. [ABSTRACT FROM AUTHOR]

Due to large line spacing and planting distances, the adoption of continuous and uniform pesticide spraying in vegetable farming can lead to pesticide waste, thus increasing cost and environmental pollution. In this paper, by applying deep learning and online identification methods, control technology for target-oriented spraying is studied with cabbages as the research object. To overcome motion blur and low average precision under strong light conditions during the operation of sprayers, an innovative YOLOV5 model implanted with a transformer module is utilized to achieve accurate online identification for cabbage fields under complex environments. Based on this concept, a new target-oriented spray system is built on an NVIDIA Jetson Xavier NX. Indoor test results show that the average precision is 96.14% and the image processing time is 51.07 ms. When motion blur occurs, the average precision for the target is 90.31%. Then, in a field experiment, when the light intensity is within the range of 3.76–12.34 wlx, the advance opening distance is less than 3.51 cm, the delay closing distance is less than 2.05 cm, and the average identification error for the cabbage diameter is less than 1.45 cm. The experimental results indicate that changes in light intensity have no significant impact on the identification effect. The average precision is 98.65%, and the savings rate reaches 54.04%. In general, the target-oriented spray system designed in this study achieves the expected experimental results and can provide technical support for field target spraying. [ABSTRACT FROM AUTHOR]

In order to integrate fragmented text data of crop disease knowledge to solve the current problems of disordered knowledge management, weak correlation and difficulty in knowledge sharing, a Chinese named-entity–relation-extraction model for crop diseases (BBCPF) was proposed in this paper by utilizing the advantage of knowledge graph in describing complex relations between disease entities in a structured form. This model was composed of two parts, i.e., named-entity recognition and relation extraction, in the form of an assembly line. To deal with the different meanings of Chinese crop disease terms in different contexts and to better obtain the contextual information, the BERT model was introduced for dynamic vector representations. Then, the BiLSTM layer was used to learn long-distance text information, and the CRF was applied to obtain the globally optimal labeling sequence, so as to output the crop disease entities. According to the entity category, the entities were divided as subjects and objects, which were then input into the disordered language model PERT to extract the contextual features of the relation data. At last, the fully connected layer was used to decode the information and output the crop disease entity-relation triples. The experiment results show that, on the self-built disease corpus dataset, the Precision, Recall, and F1-Score values of the established model reached 85.63%, 79.46% and 82.43%, respectively, for entity extraction, and reached 97.96%, 98.43% and 98.16%, respectively, for relation extraction. This paper provides an effective method for information extraction in the construction of Chinese crop disease domain knowledge graphs. [ABSTRACT FROM AUTHOR]

A series of Ca-doped yttrium iron garnet ceramics (Y 3- x Ca x Fe 5 O 12 (x = 0, 0.05, 0.1, 0.15, 0.2)) were prepared by the solid state method. The structure, morphology, valence fluctuation, thermal analysis, magnetization, magnetodielectric and Mössbauer spectrum of Y 3- x Ca x Fe 5 O 12 ceramics were systematically studied. All of the samples crystalized as single phase of garnet structure with the grain sizes of 2–3 μm. The temperature dependent percentage weight loss and heat flow were investigated by thermogravimetric (TG) and differential scanning caborimetry (DSC). The structure of Y 3- x Ca x Fe 5 O 12 ceramics was studied by XRD, Raman and Infrared (IR) spectra. The results indicated that Ca2+ substitution leads to lattice expansion and presence of Ca–O octahedron, which causes the concentration of Fe2+ decrease and Fe4+ increase. As a result, the saturation magnetization first increases and then decreases with the increase of Ca concentration. A six lines asymmetry of magnetic components were observed in Mössbauer spectrum, which was fitted by two hyperfine magnetic sextets corresponding to tetrahedral and octahedral sites, respectively. With the increase of Ca concentration, the intrinsic magnetodielectric (MD) effect disappears completely due to the dramatic decrease of Fe2+-Fe3+ dipoles. Fe3+-Fe4+ dipoles do not contribute to the MD properties of Y 3- x Ca x Fe 5 O 12 ceramics due to the lack of the defects such as electrons or holes being used as media. • Ca2+ doping in YIG ceramics results in the nonmonotonic variation of its saturation magnetization. • Intrinsic MD effect are observed at room temperature and its value decreases with the increase of Ca2+ doping. • The decrease of Fe2+-Fe3+ dipoles results in the gradually decrease and ultimate vanishing of the MD effect. [ABSTRACT FROM AUTHOR]

To improve the production efficiency and reduce the labor cost of seedling operations, cabbage was selected as the research subject, and a novel approach based on the attention mechanism combining the deep convolutional neural network (DCNN) and long short-term memory (LSTM) is proposed. First, the cabbage growth data and environmental monitoring data were normalized, and input samples were obtained by sliding the time window. Then, the DCNN and the LSTM were used to extract the spatial feature information and temporal correlation of the samples, respectively. At the same time, the attention mechanism was used to set the weight coefficients of different feature information and highlight the role of the main features of the sample in the model, thereby improving the prediction accuracy. By analyzing the experimental data collected by the Shandong Seedling Plant, the DCNN-LSTM method based on the proposed attention mechanism achieved good prediction results, providing experience for the engineering application of decision-making regarding seedling transplanting time. The experimental data showed that the mean absolute error, root-mean-square error, mean absolute percentage error, and symmetric mean absolute percentage error of the prediction results of this method were 0.356, 0.507, 0.157, and 0.082, respectively. Compared with the CNN, LSTM, LSTM-Attention and CNN-LSTM models, this model showed higher prediction accuracy. [ABSTRACT FROM AUTHOR]

Neutron-focusing mirrors with both a nested structure and supermirror coatings enable a large collecting area, and consequently a large neutron current gain for mirror-based instruments. A nested neutron-focusing supermirror assembly has been fabricated with nested conical optics. Here, we demonstrate a series of small-angle neutron scattering (SANS) measurements using this supermirror assembly. A prototype instrument was built to measure the absolute scattering intensity of a solution of silica colloids, and the instrumental resolution was calibrated with highly ordered SBA-15. By comparing the theoretical and experimental values of the minimum Q (Q min), the non-geometrical component leading to the deterioration in Q min is estimated and employed to calculate the performance of mirror-based SANS instruments. These results, together with the discussion on the capability of mirror-based instruments and other practical issues, are meaningful for promoting the development of neutron-focusing optics and mirror-based SANS instruments. [ABSTRACT FROM AUTHOR]

Papaver somniferum (opium poppy) is not only a source of raw material for the production of medical narcotic analgesics but also the major raw material for certain psychotropic drugs. Therefore, it is stipulated by law that the cultivation of Papaver somniferum must be authorized by the government under stringent supervision. In certain areas, unauthorized and illicit Papaver somniferum cultivation on private-owned lands occurs from time to time. These illegal Papaver somniferum cultivation sites are dispersedly-distributed and highly-concealed, therefore becoming a tough problem for government supervision. The low-altitude inspection of Papaver somniferum cultivation by unmanned aerial vehicles has the advantages of high efficiency and time saving, but the large amount of image data collected needs to be manually screened, which not only consumes a lot of manpower and material resources but also easily causes omissions. In response to the above problems, this paper proposed a two-stage (target detection and image classification) method for the detection of Papaver somniferum cultivation sites. In the first stage, the YOLOv5s algorithm was used to detect Papaver somniferum images for the purpose of identifying all the suspicious Papaver somniferum images from the original data. In the second stage, the DenseNet121 network was used to classify the detection results from the first stage, so as to exclude the targets other than Papaver somniferum and retain the images containing Papaver somniferum only. For the first stage, YOLOv5s achieved the best overall performance among mainstream target detection models, with a Precision of 97.7%, Recall of 94.9%, and mAP of 97.4%. For the second stage, DenseNet121 with pre-training achieved the best overall performance, with a classification accuracy of 97.33% and a Precision of 95.81%. The experimental comparison results between the one-stage method and the two-stage method suggest that the Recall of the two methods remained the same, but the two-stage method reduced the number of falsely detected images by 73.88%, which greatly reduces the workload for subsequent manual screening of remote sensing Papaver somniferum images. The achievement of this paper provides an effective technical means to solve the problem in the supervision of illicit Papaver somniferum cultivation. [ABSTRACT FROM AUTHOR]

In order to improve the management efficiency of peach orchards, this paper considers the cooperative operation scheme of multiple unmanned tractors. According to the actual situation, this paper constructs the path planning model of multiple unmanned tractors in a standard peach orchard, designs the objective function to optimize the total turning time and total operating time according to the tractor driving parameters, and solves it by improving the differential evolution algorithm. Aiming at the premature convergence problem, the permutation matrix is introduced to represent the driving paths of multiple unmanned tractors. Then, the dynamic parameters are adopted to make the parameters change with the number of iterations, and the elite selection strategy is used to eliminate the redundant feasible solutions. An Adaptive Elite Differential Evolution (AEDE) algorithm suitable for multi-tractor path optimization is proposed. The results show that, compared with the traditional Differential Evolution algorithm (Differential Evolution, DE), the total turning time and total operating time in the rectangular peach orchard optimized by AEDE are reduced by 3.34% and 0.87%, respectively. Compared with the block operation, the total turning time and total operating time of the AEDE-optimized rectangular peach orchard operation path were reduced by 37.37% and 9.47%, respectively. Experiments show that AEDE, which optimizes the operating path of multi-tractors in standard peach orchards, is able to improve the efficiency and reduce the operating time. [ABSTRACT FROM AUTHOR]

With the development of modern agricultural facilities, crop diseases recognition, nutritional status and morphology achieved rapid growth. To avoid yield loss caused by the delay of disease detection, digital images that contain information with respect to crop growth, disease type and nutrition deficiency have been studied by some researchers. However, traditional image processing methods fail to extract typical disease features of crop images with ambiguous disease information. In this paper, a crop disease image recognition technique based on the salient region and hierarchical indexing was proposed. Improved Harris algorithm and maximum radius were used to calculate the widest salient region. In order to eliminate the effect of different salience distribution ranges between different features, a group of images in the cucumber disease image library were normalized. Experiment results indicate that the time complexity of each algorithm will go up as the size of the dataset increase. Especially when testing large datasets, nonhierarchical and nonclustering, hierarchical and nonclustering and hierarchical based on points all tend to raise the algorithm's time complexity. Plant Village dataset and AI Challenger 2018 dataset were utilized to compare the recognition performances among the models based on machine learning, neural network, deep learning and our methods. The experiment results show that the method proposed in this paper is capable of recognizing local similar images effectively rather than global similar images, therefore, it has better recognition performance than the model learning methods in the early detection stage of crop disease. [ABSTRACT FROM AUTHOR]

To enhance the transplantation effectiveness of vegetables and promptly formulate subsequent work strategies, it is imperative to study the recognition approach for transplanted seedlings. In the natural and complex environment, factors like background and sunlight often hinder accurate target recognition. To overcome these challenges, this study explores a lightweight yet efficient algorithm for recognizing cabbage transplantation states in natural settings. Initially, FasterNet is integrated as the backbone network in the YOLOv5s model, aiming to expedite convergence speed and bolster feature extraction capabilities. Secondly, the introduction of the GAM attention mechanism enhances the algorithm's focus on cabbage seedlings. EIoU loss is incorporated to improve both network convergence speed and localization precision. Lastly, the model incorporates deformable convolution DCNV3, which further optimizes model parameters and attains a superior balance between accuracy and speed. Upon testing the refined YOLOv5s target detection algorithm, improvements were evident. When compared to the original model, the mean average precision (mAP) rose by 3.5 percentage points, recall increased by 1.7 percentage points, and detection speed witnessed an impressive boost of 52 FPS. This enhanced algorithm not only reduces model complexity but also elevates network performance. The method is expected to streamline transplantation quality measurements, minimize time and labor inputs, and elevate field transplantation quality surveys' automation levels. [ABSTRACT FROM AUTHOR]

To overcome the limitations of traditional data collection methods in large-scale farmland wireless sensor network, in this study, we introduce a mobile sink and propose a virtual potential field-based strategy for mobile sink path planning. Virtual potential field-based strategy constructs a virtual field based on the residual energy, data generation rate, location information and cache urgency of nodes in the monitoring area. The stronger the virtual field, the more attractive it will be to mobile sink, which consequently affects the mobile path of sink node. Rendezvous points are selected in accordance with the maximum-farthest criterion, and the shortest path connecting all rendezvous points is taken as the mobile path of sink. Furthermore, the monitoring nodes employ the distance probability transmission strategy to have the transmission moment selected and the energy consumption optimized with reference to the path control information sent by the sink. The virtual potential fields and the rendezvous points are recalculated periodically according to the dynamic changes of both the node residual energy and the real-time cache. The simulation results showed that excellent transmission efficiency and network lifetime, and the combination of virtual potential field-based strategy and distance probability transmission strategy can have the fairness and real time of nodes guaranteed, thus it may meet the needs of large-scale farmland data collection. [ABSTRACT FROM AUTHOR]

• A two-stage dense captioning model for vegetable leaf diseases is proposed. • The model can identify diseases and explain the results. • Mixed disease recognition problem solved. • The recognition ability of small lesion in the model is improved. The latest advances in deep learning technology make it possible to recognize vegetable diseases from leaf images. The existing disease recognition methods based on computer vision have shown exciting achievements in terms of accuracy, stability, and portability. However, these methods cannot provide a decision-making basis for the final results, and lack a text basis to support the users' judgement. Disease diagnosis is a risky decision. If the detection method lacks transparency, the users will not be able to fully trust the recognition results, which greatly limits the application of various recognition methods based on deep learning. Aiming at the problem of low "man–machine" credibility due to the fact that deep learning-based methods are unable to provide decision-making basis, this paper proposed a two-stage image dense captioning model named "DFYOLOv5m-M2Transformer", which can generate description sentences of visualized disease features on the basis of the recognized diseased area. Firstly, we established a target detection dataset and a dense captioning dataset containing leaf images of 10 diseases, involving 2 vegetables, i.e., cucumber and tomato. Secondly, we chose the DFYOLOv5m network as the disease detector to extract the diseased area from the image, and the M2-Transformer network as the decision basis generator to generate description sentences of disease features. Then, the Bi-Level Routing Attention module was introduced to extract fine-grained features under complex backgrounds in order to resolve the problem of poor feature extraction in case of mixed diseases. Finally, we used Atrous Convolution to expand the receptive field of the model, and fused NWD and CIoU to improve the model's performance in detecting small targets. The experimental results show that the IoU and Meteor joint evaluation indicator of DFYOLOv5m-M2Transformer achieved a mean Average Precision (mAP) of 94.7 % on the dense captioning dataset, which was 7.2 % higher than that of the best-performing model Veg-DenseCap in the control group. Moreover, the decision basis that is automatically generated by the proposed model is characterized by the advantages of high accuracy, correct grammar and large sentence variety. The outcome of this study provides a new idea for optimizing the user experience in using vegetable disease recognition models. [ABSTRACT FROM AUTHOR]

Abstract Polycrystalline Bi 4 (Ti 3- x Fe x /2 Co x /2)O 12 (0 ≤ x ≤ 0.4) ceramics are prepared by solid state reaction method. The low temperature ferromagnetism is successfully induced by Fe/Co co-doping. Particularly, the x = 0.3 ceramics possess the largest remnant magnetization (2 M r) of about 0.272 emu/g at 100 K. Furthermore, all the doped ceramics undergo a ferromagnetic to paramagnetic transition, and show typical spin glass behavior, which is confirmed by the isothermal remnant magnetization and Memory Effect. The magnetic mechanisms probably originate from the coexistence and competition between ferromagnetic double exchange interactions and antiferromagnetic superexchange interactions, which lead to the appearance of the spin glass behavior in the doped ceramics and the strongest ferromagnetism in the x = 0.3 ceramics. The present work provides basic data and theoretical reference for the exploitation of the new multifunctional materials. [ABSTRACT FROM AUTHOR]

Neutron radiography is a powerful method of probing the structure of materials based on attenuation of neutrons. This method is most suitable for materials containing heavy metals, which are not transparent to X-rays, for example irradiated nuclear fuel and other nuclear materials. Neutron radiography is one of the first non-distractive post-irradiated examination methods, which is applied to gain an overview of the integrity of irradiated nuclear fuel and other nuclear materials. However, very powerful gamma radiation emitted by the samples is damaging to the electronics of digital imaging detectors and has so far precluded the use of modern detectors. Here we describe a design of a neutron microscope based on focusing mirrors suitable for thermal neutrons. As in optical microscopes, the sample is separated from the detector, decreasing the effect of gamma radiation. In addition, the application of mirrors would result in a thirty-fold gain in flux and a resolution of better than 40 μ m for a field-of-view of about 2.5 cm. Such a thermal neutron microscope can be useful for other applications of neutron radiography, where thermal neutrons are advantageous. [ABSTRACT FROM AUTHOR]

• Real time detection of strawberry leaves powdery mildew using deep neural network. • DW-CSP backbone and DW-Neck were proposed to replace the backbone and neck of YOLOv4. • 4 New types of network structure based on YOLOv4 were proposed. • DAC-YOLOv4 is faster and more accuracy than YOLOv4. • A method of powdery mildew disease index of strawberry leaves was proposed. Strawberry powdery mildew (PM) is the main disease affecting the yield and quality of strawberries in recent years, which always appears on the back side of leaves in the early stage. Traditional methods of disease detection are labor-intensive and time-consuming. In this paper, we proposed a computer vision algorithm for strawberry leaf PM and infected leaves (IL) detection in complex background. Then we additionally proposed the estimation index of strawberry leaf PM for disease assessment. The original YOLOv4 backbone and neck are replaced by the proposed backbone and neck with depthwise convolution and hybrid attention mechanism, and the improvement can be made to decrease the size of the model and retain the performance. By combining the proposed backbone and neck, four new network structures are designed and evaluated, and the best one was named DAC-YOLOv4. Compared with YOLOv4, the mean average precision (mAP) of DAC-YOLOv4 reaches 72.7%, while the size is greatly compressed. To confirm the effectiveness of the proposed model, we compare DAC-YOLOv4 with five algorithms, and experimentally show that DAC-YOLOv4 performs well. We also deploy the algorithm on the Jetson Xavier NX and Jetson Nano, and the speed of DAC-YOLOv4 is 43 and 20 FPS, respectively, which can meet the real-time detection requirements. In summary, the experimental results indicate that the DAC-YOLOv4 proposed in this paper has good performance in strawberry leaf PM detection on the embedded platform, and the method to attain the disease index provides a solution for the early detection and prevention of strawberry PM. [ABSTRACT FROM AUTHOR]

The development of resistant cucumber varieties is of a great importance for reducing the production loss caused by root-knot nematodes. After cucumber plants are infected with root-knot nematodes, their roots will swell into spherical bumps. Rapid and accurate detection of the infected sites and assessment of the disease severity play a key role in selecting resistant cucumber varieties. Because the locations and sizes of the spherical bumps formed after different degrees of infection are random, the currently available detection and counting methods based on manual operation are extremely time-consuming and labor-intensive, and are prone to human error. In response to these problems, this paper proposes a cucumber root-knot nematode detection model based on the modified YOLOv5s model (i.e., YOLOv5-CMS) in order to support the breeding of resistant cucumber varieties. In the proposed model, the dual attention module (CBAM-CA) was adopted to enhance the model's ability of extracting key features, the K-means++ clustering algorithm was applied to optimize the selection of the initial cluster center, which effectively improved the model's performance, and a novel bounding box regression loss function (SIoU) was used to fuse the direction information between the ground-truth box and the predicted box so as to improve the detection precision. The experiment results show that the recall (R) and mAP of the YOLOv5s-CMS model were improved by 3% and 3.1%, respectively, compared to the original YOLOv5s model, which means it can achieve a better performance in cucumber root-knot nematode detection. This study provides an effective method for obtaining more intuitive and accurate data sources during the breeding of cucumber varieties resistant to root-knot nematode. [ABSTRACT FROM AUTHOR]

Natural resources, such as light and CO 2 , are required to increase crop yields in protected agriculture. Developing efficient regulation methods to reduce costs and increase yields has become a major challenge for actual production. This work proposed a framework for meeting this challenge and supporting the management of environmental regulation. Photosynthesis rate was used as the yield index, and a photosynthesis rate model was constructed using support vector regression. Moreover, a regulation cost function for light and CO 2 was constructed. Then, the non-inferior solutions (NIS) for environmental regulation were determined by an improved non-dominated sorting genetic algorithm Ⅱ (i-nsGA Ⅱ). The curvature theory was applied to calculate the optimal solution in the NIS set. A set of seedling tomato data was used to demonstrate the usefulness of the proposed method. Compared with the other two regulation methods (saturation regulation method and U-chord method), the results indicated that the proposed method had the best performance. About 80% of the maximum photosynthesis rate could be obtained at a regulation cost of about 51%. This research presents a new strategy for optimizing the greenhouse environment, which can be applied as a knowledge-informed expert system to manage environmental regulation. [ABSTRACT FROM AUTHOR]

• A cucumber leaf disease segmentation model with a two-stage architecture was proposed. • The complementary advantages of the DeepLabV3+ and U-Net models were fully exerted. • A disease severity classification method for cucumber downy mildew was proposed. • All the disease images were collected from the field with complex backgrounds. Research on the recognition and segmentation of vegetable diseases in simple environments based on deep learning has achieved a relative success. However, in complex environments, the image background often contains elements similar to the representation of leaves and disease spots, making it difficult for the recognition model to segment leaves and disease spots. Consequently, the segmentation precision is significantly reduced, which further affects the accuracy of disease severity classification. In response to this problem, while discussing and analyzing the advantages and disadvantages of DeepLabV3+ and U-Net, this study proposed a two-stage model that fuses DeepLabV3+ and U-Net for cucumber leaf disease severity classification (DUNet) in complex backgrounds. In the first stage, this model uses DeepLabV3+ to segment leaves from complex backgrounds. The images of leaves obtained after segmentation are used as the input for the second stage. In the second stage, U-Net is used to segment the diseased leaves to obtain disease spots. Finally, the ratio of the pixel area of disease spots over the pixel area of leaves is calculated so as to classify the disease severity. The experiment results show that the proposed model is able to segment leaves and disease spots from complex backgrounds in a step-by-step manner so as to complete disease severity classification. The leaf segmentation accuracy reached 93.27%, the Dice coefficient of disease spot segmentation reached 0.6914, and the average disease severity classification accuracy reached 92.85%. Compared with other models, the model proposed in this study has higher robustness, segmentation precision and classification accuracy, providing important ideas and methods for classifying the severity of cucumber leaf diseases in complex backgrounds. [ABSTRACT FROM AUTHOR]

In the question-and-answer (Q&A) communities of the "China Agricultural Technology Extension Information Platform", thousands of rice-related Chinese questions are newly added every day. The rapid detection of the same semantic question is the key to the success of a rice-related intelligent Q&A system. To allow the fast and automatic detection of the same semantic rice-related questions, we propose a new method based on the Coattention-DenseGRU (Gated Recurrent Unit). According to the rice-related question characteristics, we applied word2vec with the TF-IDF (Term Frequency–Inverse Document Frequency) method to process and analyze the text data and compare it with the Word2vec, GloVe, and TF-IDF methods. Combined with the agricultural word segmentation dictionary, we applied Word2vec with the TF-IDF method, effectively solving the problem of high dimension and sparse data in the rice-related text. Each network layer employed the connection information of features and all previous recursive layers' hidden features. To alleviate the problem of feature vector size increasing due to dense splicing, an autoencoder was used after dense concatenation. The experimental results show that rice-related question similarity matching based on Coattention-DenseGRU can improve the utilization of text features, reduce the loss of features, and achieve fast and accurate similarity matching of the rice-related question dataset. The precision and F1 values of the proposed model were 96.3% and 96.9%, respectively. Compared with seven other kinds of question similarity matching models, we present a new state-of-the-art method with our rice-related question dataset. [ABSTRACT FROM AUTHOR]

Poppy is a special medicinal plant. Its cultivation requires legal approval and strict supervision. Unauthorized cultivation of opium poppy is forbidden. Low-altitude inspection of poppy illegal cultivation through unmanned aerial vehicle is featured with the advantages of time-saving and high efficiency. However, a large amount of inspection image data collected need to be manually screened and analyzed. This process not only consumes a lot of manpower and material resources, but is also subjected to omissions and errors. In response to such a problem, this paper proposed an inspection method by adding a larger-scale detection box on the basis of the original YOLOv3 algorithm to improve the accuracy of small target detection. Specifically, ResNeXt group convolution was utilized to reduce the number of model parameters, and an ASPP module was added before the small-scale detection box to improve the model's ability to extract local features and obtain contextual information. The test results on a self-created dataset showed that: the mAP (mean average precision) indicator of the Global Multiscale-YOLOv3 model was 0.44% higher than that of the YOLOv3 (MobileNet) algorithm; the total number of parameters of the proposed model was only 13.75% of that of the original YOLOv3 model and 35.04% of that of the lightweight network YOLOv3 (MobileNet). Overall, the Global Multiscale-YOLOv3 model had a reduced number of parameters and increased recognition accuracy. It provides technical support for the rapid and accurate image processing in low-altitude remote sensing poppy inspection. [ABSTRACT FROM AUTHOR]

• Fe 1-x S/SC nanosheets derived from Fe-MOFs as anode for PIBs. • The Fe 1-x S/SC nanosheets promote the penetration of electrolyte. • This special structure accelerates the transport speed of K+. • Sulfur doping increase more active sites and promote the storage of K+. • Fe 1-x S/SC electrode delivers superior rate and outstanding cycle performance. Metal sulfides are popular in energy storage system for unique advantages, including high theoretical capacity and readily available raw materials. However, there are some obstacles to overcome before application in potassium ion batteries (PIBs). Huge changes in volume during insertion/extraction of K+, which will cause the structure destruction of the electrode and lead to rapid decay of capacity. In addition, due to the poor electrical conductivity of Fe 1-x S, the rate performance is always poor. Herein, Fe 1-x S/S-doped carbon nanosheets (Fe 1-x S/SC) were synthesized. When applied Fe 1-x S/SC as anode for PIBs, it exhibits excellent electrochemical performance: a reversible capacity of 430 mAh g−1 can be maintained at 50 mA g−1 after 300 cycles, a super rate performance of 255 and 170 mAh g−1 at 1000 and 2000 mA g−1 can be observed. [ABSTRACT FROM AUTHOR]

• Proposed a crop-specific control method for soil moisture and light intensity. • Analyzed the contribution of interaction terms by the generalized additive model. • Adopted the gaussian curvature algorithm to analyze the variation of the surface. • Established a regulation strategy based on the Gaussian curvature maximum. Soil moisture (SM) exerts a significant impact on crop growth, interacting with environmental factors such as temperature, photosynthetic photon flux density (PPFD), and CO 2 , ultimately affecting crop photosynthesis (Pn). This study employs a nested experimental design to investigate the photosynthetic activity of cucumber seedlings under diverse environmental conditions and establishes a support vector regression (SVR) model for Pn prediction. The SVR model takes temperature, PPFD, SM, and CO 2 concentration as inputs and demonstrates a high level of accuracy (The model's coefficient of determination = 0.9830, root mean square error = 0.9138). Subsequently, through a generalized additive model, the study unveils the significant impact of interactions between SM and PPFD on Pn. Accordingly, this research constructs a Pn response surface based on these two factors and identifies the maximum point of Gaussian curvature on this surface. Polynomial regression is applied to these points, yielding a comprehensive regulation strategy for SM and PPFD. In comparison to traditional methods based on maximizing Pn, this innovative approach reduces Pn by 12.9 % while significantly conserving light (35.59 %) and water (32.80 %) consumption. Although no significant changes are observed in crop physiological traits (plant height, stem diameter, dry weight, fresh weight), substantial variations are noted in irrigation volume and PPFD consumption. Thus, the regulation strategy proposed in this study embodies efficiency and energy conservation in greenhouse crop cultivation. [ABSTRACT FROM AUTHOR]

• An environmental control method aims at high photosynthesis rate and low control cost. • Discrete surface curvature measuring the curved degree of surfaces was proposed. • Non-dominated sorting genetic algorithm Ⅱ was improved. • Constraint region of photosynthetic response surface was defined. Protected greenhouse provides a suitable micro-environment for crop growth, which enhances the quality and yield of agricultural products. How to optimize the greenhouse environment is currently a pressing issue in protected agriculture. In this research, we propose a new environmental control method from the perspectives of crop photosynthesis rate and environmental control cost. The photosynthesis rate data are collected from a nested experiment of tomatoes, and these data are combined with the support vector regression algorithm to establish a photosynthesis rate model. The cost function of greenhouse environmental control is established by theoretical calculation. The non-dominated solution sets of photosynthesis rate and control cost are obtained through an improved non-dominated sorting genetic algorithm Ⅱ. To determine the optimal solution among the numerous Pareto solutions, the geometric characteristics of the photosynthetic response surface are considered to constrain the control target. Then, a discrete surface curvature method is proposed to computer the constraint region. The intersection between the Pareto solutions and the constraint region is regarded as the environmental control target. The validation experiment demonstrates that the proposed method can effectively enhance photosynthetic accumulation and reduce control cost. Compared to the traditional saturation method, this method achieves a 53% cost reduction while still obtaining 89% photosynthetic accumulation. Compared to the U-chord method, this method increases photosynthetic accumulation by 12% with only a 3% increase in cost. The research has important implications for protected agriculture production. [ABSTRACT FROM AUTHOR]

For the rapid and accurate identification of cow reproduction and healthy behavior from mass surveillance video, in this study, 400 head of young cows and lactating cows were taken as the research object and analyzed cow behavior from the dairy activity area and milk hall ramp. The method of object recognition based on image entropy was proposed, aiming at the identification of motional cow object behavior against a complex background. Calculating a minimum bounding box and contour mapping were used for the real-time capture of rutting span behavior and hoof or back characteristics. Then, by combining the continuous image characteristics and movement of cows for 7 d, the method could quickly distinguish abnormal behavior of dairy cows from healthy reproduction, improving the accuracy of the identification of characteristics of dairy cows. Cow behavior recognition based on image analysis and activities was proposed to capture abnormal behavior that has harmful effects on healthy reproduction and to improve the accuracy of cow behavior identification. The experimental results showed that, through target detection, classification and recognition, the recognition rates of hoof disease and heat in the reproduction and health of dairy cows were greater than 80%, and the false negative rates of oestrus and hoof disease were 3.28% and 5.32%, respectively. This method can enhance the real-time monitoring of cows, save time and improve the management efficiency of large-scale farming. [ABSTRACT FROM AUTHOR]

We conducted a comparative study on structural, magnetic and electric properties of MnTiO 3 ceramics sintered in air (MTO-A) and nitrogen (MTO-N) to explore the impact of oxygen vacancies (OVs) on them. The X-ray diffraction patterns show that MTO-N possesses a lattice contraction induced by OVs comparing with MTO-A. Distinct spin glass (SG) with T sg (41 K) and magnetic exchange bias (EB) are only observed in the MTO-N sample. The results indicate that OVs may well lead to the enhanced magnetization, thereby generating SG behavior. Moreover, the dielectric characteristics of MTO-A and MTO-N are investigated additionally, two disparate relaxation mechanisms of Maxwell–Wagner relaxation and OVs migration are illuminated respectively, which further support the vital function of OVs. [ABSTRACT FROM AUTHOR]

• A predictive model of photosynthetic rate was proposed. • U-chord curvature was used to determine knee point of a light-response curve. • A decision-making model of light environment control was proposed for tomato. • The decision-making model could improve Pn with a few energy consumptions. Light, the energy source for crop photosynthesis, is a key factor for plant growth. The present study proposes a decision-making model of light environment control. The photosynthesis rate of tomato seedlings under different light intensities, temperatures, and CO 2 concentrations was determined in a nested experiment. These data were used to construct a predictive model of the photosynthesis rate using the support vector regression method, with an R2 of 0.9862, a root mean square error of 1.39 μmol·m−2·s−1, and a mean absolute error of 1.18 μmol·m−2·s−1. In total, 861 discrete light-response curves were obtained based on the predictive model, and their knee points were computed using the U-chord curvature method. These knee points were used to form a dataset for constructing a decision-making model for light environment control, with an R2 of 0.984 and a root mean square error of 9.55 μmol·m−2·s−1. The results of the validation experiment suggested that the average relative error of the model was 1.92%, indicating the robustness of the model. Compared with those of the light saturation control method, the average light demand for the decision-making model decreased by 60.49%, whereas the average photosynthesis rate reduced by 24.40%. Although the photosynthesis rate lost a bit, the rate of light saving is almost three times more than the rate of photosynthesis rate decreased slightly, which improved the production efficiency of tomato. [ABSTRACT FROM AUTHOR]

• This article propose a dual-branch, efficient, channel attention (DECA) module which uses a dual-branch 1D convolution operation to filter effective feature information. • It also introduces an adaptive convolution kernel parameter k and an adaptive parameter α to participate in the reverse training of the model. • The DECA_ResNet model is proposed to realize crop disease recognition. • We verified the DECA_ResNet model on the AI Challenger 2018 dataset, PlantVillage dataset, and self-collected cucumber disease dataset; the disease recognition accuracies are 86.35%, 99.74% and 98.54%, respectively. Efficient and accurate recognition of crop diseases plays an important role in disease prevention. Aiming at the low accuracy of existing crop disease recognition methods, we improve the attention mechanism and residual neural network (ResNet) to propose a dual-branch, efficient, channel attention (DECA)-based crop disease recognition model. The DECA module uses a dual-branch 1D convolution operation to filter effective feature information. It also introduces an adaptive convolution kernel parameter k and an adaptive parameter α to participate in the reverse training of the model so that the model can independently select effective features and establish dependencies between channels. Then, the DECA module is added to the residual module to recalibrate the channel characteristics and improve the characteristic representability of the residual module. Finally, a DECA_ResNet model is proposed to realize crop disease recognition. The crop disease model based on improved attention is verified on the AI Challenger 2018 dataset, PlantVillage dataset, and self-collected cucumber disease dataset; the disease recognition accuracies are 86.35%, 99.74% and 98.54%, respectively, which are higher than those of the ResNet and squeeze-and-excitation network (SENet) before improvement. The recognition accuracy of the proposed model in the PlantVillage dataset is higher than that of existing models. The experimental results show that the proposed plant recognition model with 18 layers (DECA_ResNet18) can achieve a high recognition accuracy. The DECA module can independently select each branch feature without causing a significant increase in the number of parameters, which improves the plant disease recognition accuracy and reduces the extraction of redundant features. [ABSTRACT FROM AUTHOR]

• An image text multimodal collaborative representation recognition model was proposed. • A knowledge graph for specific crop diseases was established. • Knowledge aided disease identification. • The recognition results were interpreted in several ways. Identification methods of crop diseases based on image modality alone have achieved relative success under limited and restricted conditions. As a data-driven technology, its performance depends on a large amount of image labeling data. Many of the existing methods neglected the role and value of other modal data except images and only relies on low-level image features for disease recognition without utilizing high-level domain knowledge, leading to poor credibility and interpretability of identification results. This paper targets tomato and cucumber common invasive diseases recognition as the research object. First, for the problem of insufficient utilization of multimodal data in existing models, the semantic embedding methods for disease images and disease description texts were examined, and the correlation and complementarity between the two types of modal data was utilized to realize the joint representation learning of disease features. Second, in response to the requirements of reliable identification and interpretability of diseases, the knowledge representation and knowledge embedding mechanism in the field of disease identification was studied, and the high-level domain knowledge graph was used as the external guidance for image feature learning and disease identification. Lastly, a disease identification model based on "image-text" multimodal collaborative representation and knowledge assistance (ITK-Net) was constructed. The proposed model achieved an identification accuracy, precision, sensitivity and specificity of 99.63%, 99%, 99.07% and 99.78% respectively on a dataset composed of "image-text" pairs. Meanwhile, semantic interpretation was performed on the model inference process. The achievement of this paper can offer a new method for disease identification based on multimodal data and domain knowledge, which might help improve the intelligence level of crop disease identification. [ABSTRACT FROM AUTHOR]

• Images with different complexity backgrounds were collected. To measure the detection performance on various images, a method was proposed to quantify image background complexity. • We constructed an efficient detection model EFDet to detect cucumber disease in complex images. • EFDet has high robustness for complex environments and fewer parameters and computation. Improving the application capability of the disease detection model is a key issue in the field of agricultural informatization. The complex backgrounds, image diversity, and model complexity are the main factors that affect the realization of automatic disease recognition. This study constructs an efficient detection model (EFDet), which mainly consists of the efficient backbone network, a feature fusion module, and a predictor. EFDet improves the detection effect for cucumber leaves in complex backgrounds by fusing feature maps at different levels. We collected three category cucumber leaves including downy mildew, bacterial angular spot, and health to construct the cucumber disease dataset. It contains 7,488 images with three complexity levels for model training and evaluation. YOLO V3-V5, EfficientDet-D1, YOLO V3-ASFF, and other six detection models as the comparison models, we verify the EFDet performance in terms of model size , FLOPs , and mAP. Experimental results show that EFDet has strong robustness for cucumber disease leaf in complex environments. It also has smaller parameters and calculations that are suitable for actual applications. [ABSTRACT FROM AUTHOR]

This paper introduces a series of experiments with an ALBERT over match-LSTM network on the top of pre-trained word vectors, for accurate classification of intelligent question answering and thus the guarantee of precise information service. To improve the performance of data classification, a short text classification method based on an ALBERT and match-LSTM model was proposed to overcome the limitations of the classification process, such as few vocabularies, sparse features, large amount of data, lots of noise and poor normalization. In the model, Jieba word segmentation tools and agricultural dictionary were selected to text segmentation, GloVe algorithm was then adopted to expand the text characteristic and weighted word vector according to the text of key vector, bi-directional gated recurrent unit was applied to catch the context feature information and multi-convolutional neural networks were finally established to gain local multidimensional characteristics of text. Batch normalization, Dropout, Global Average Pooling and Global Max Pooling were utilized to solve overfitting problem. The results showed that the model could classify questions accurately, with a precision of 96.8%. Compared with other classification models, such as multi-SVM model and CNN model, ALBERT+match-LSTM had obvious advantages in classification performance in intelligent Agri-tech information service. [ABSTRACT FROM AUTHOR]

• A resource-saving environmental control method is proposed for crops. • U-chord curvature algorithm was adopted to determine knee points of a curve. • Optimal regulation space is determined based on knee points. • The mathematical and physiological significance of knee points was discussed. Temperature, light and CO 2 are three environmental factors that directly affect the photosynthetic rate of plants. Exploring the relationship between these three factors and the photosynthetic rate, and optimizing the environmental conditions, is the key to realize efficient production of greenhouse crops. A nested experiment was carried out to measure the photosynthetic rate of cucumber seedlings under different environmental conditions. On the basis of these data, a photosynthetic rate prediction model was established by using machine learning method. A method that avoiding the excessive consumption of light and CO 2 resources and promoting the photosynthetic rate based on U-chord curvature is proposed for more efficient regulation of photosynthetic rate. Under different temperature conditions, the prediction model can be used to construct photosynthetic rate surfaces under the interaction of light and CO 2. After the surface is discretized, the maximum U-chord curvature points of discrete light response and CO 2 response curves are calculated. These points are fitted by polynomials as the regulation boundary to obtain the target space that can achieve efficient regulation. The prediction model, with temperature, light and CO 2 as inputs, had high accuracy, and the regulation method was effective. The determination coefficient and root mean square error of the prediction model were 0.99 and 0.85 µmol·m−2·s−1, respectively. Compared with the traditional maximum photosynthetic rate regulation method, this new method reduced the photosynthetic rate by under 16%, but saved 41% of light and 49% of CO 2 inputs, which illustrated this method improved production efficiency while basically maintained maximum photosynthetic rate. [ABSTRACT FROM AUTHOR]

• A novel model PRP-Net for vegetables disease recognition in complex background is proposed. • A framework combines regional recommendation network and progressive learning network. • Accuracy improved by 13% using the proposed approach. • All the disease images were collected from the field environment. The method of vegetable disease recognition based on deep learning has achieved relative success under limited conditions, but a recognition model targeting at simple backgrounds may be subjected to greatly-reduced accuracy in actual production environments. This is due to the fact that the disease images obtained from actual production environments are usually of complex backgrounds, which may contain elements similar to the disease features or symptoms. Such a phenomenon significantly increases the difficulty of disease recognition. In response to this problem, we proposed a vegetable disease recognition model for complex backgrounds based on region proposal and progressive learning (PRP-Net). This model can locate the regions of interest in the images of diseased leaves based on the region proposal network. The located regions will then be cropped and enlarged as the input to proceed with progressive learning in a finer-scale network. The region proposal part of this model can guide the model to focus on the regions of interest in the disease images with complex backgrounds in a weakly supervised manner, so as to avoid the expensive costs to manually label the key regions of the image. The progressive learning network allows the model to learn global features and local delicate features in a progressive manner. In a self-collected image dataset containing 6 types of complex-background vegetable diseases, this model achieves an average recognition accuracy of 98.26%, which is 4.46 percentage points higher than that of the original region proposal network framework (RA-CNN). Meanwhile, it is also superior to any feature extraction network that is used alone. This study provides useful ideas and methodological concepts for recognizing vegetable diseases under complex backgrounds. [ABSTRACT FROM AUTHOR]

• A model based on image text collaborative representation learning was proposed. • Text improves the learning ability of the model on small samples. • The back leaf image is added to the model to support the disease diagnosis process. • All the disease images were collected from the field environment. Automatic recognition of vegetable diseases in complex backgrounds is an urgent need in the field of agricultural informatization. The recognition methods based on deep learning have achieved excellent performance in disease diagnosis and therefore have gradually become a research hotspot. However, the disease recognition models established based on deep convolutional neural networks usually need to be trained on huge disease image datasets so as to achieve an ideal outcome. Building such a kind of dataset requires a large amount of disease images and labeling information, which is often technically or economically infeasible. In this paper, a small-sample recognition model of vegetable diseases in complex backgrounds based on image text collaborative representation learning (ITC-Net) was proposed. This model combined the disease image modal information with the disease text modal information, so as to achieve collaborative recognition of disease features by utilizing the correlation and complementarity between the two types of disease information. Eventually, the ITC-Net achieved better results than either the image model or text model alone on a small dataset. To be more specific, its accuracy, precision, sensitivity and specificity are 99.48%, 98.90%, 98.78% and 99.66%, respectively. This paper proves that the multi-modal collaborative representation learning using both disease images and disease texts is an effective method to solve the problem of vegetable disease recognition in complex backgrounds with few-shot. [ABSTRACT FROM AUTHOR]

A reasonable estimation of crop water uptake depth (WUD) provides vital prerequisites for determining scientific irrigation scheduling. However, real-time acquisition of crop WUD to guide irrigation has not been reported. In the current study, a smart irrigation system based on real-time soil moisture data was developed, in which the dynamic crop WUD was estimated using the spatiotemporal characteristics of soil moisture distributions. Subsequently, the data of crop WUD were acquired using a central irrigation controller to achieve a precise irrigation depth at each irrigation event. A drip irrigation experiment for tomato (Lycopersicum esculentum) planting was conducted in a greenhouse in northern China. Wireless soil moisture sensors were installed to collect real-time soil moisture data from a 0–100 cm soil profile. Soil moisture exhibited a ladder trend for 0–60 cm and a stable trend for 60–100 cm. A regression equation (WUD i = −0.0119 d 2 + 1.9387 d - 6.5795, R 2 = 0.89) quantifying the dynamic depth of crop water uptake was successfully obtained based on the characteristics of soil moisture distribution, and a water-saving scheme (i.e. , irrigation to real-time crop WUD) was determined to guide every irrigation event via a remote automatic irrigation system. Tomato evapotranspiration (ET) calculated using soil moisture data was consistent with that calculated using indoor meteorological data, confirming the reliability of real-time soil moisture data for estimating the tomato WUD. The tomato irrigation water use efficiency (IWUE) of the established system reached 41.23 kg/m3, which was an evident improvement compared with that obtained using a traditional irrigation scheme (31.58 kg/m3). In addition, the IWUE of tomato in the established system was almost approached that of the previously published results. The findings of this study highlight the importance of reasonable water-saving irrigation scheduling and provide insights into the development of an efficient and automatic irrigation system. • Water-saving scheme embedded in the irrigation system is essential for irrigated farming. • The dynamic tomato water uptake depth is estimated from real-time soil moisture distributions. • Under- or over-irrigation is avoided by supplying water to tomato effective water uptake depth. • The developed smart irrigation system achieves more yields with less irrigation water. [ABSTRACT FROM AUTHOR]