摘要：
The photon dissipation process in photosynthetic reactions has been recognized for its biological significance in contributing to greater biosphere entropy production. While substantial progress has been made in understanding the temporal dynamics of its dissipation regulation, the spatial dynamics remain poorly understood. Here, we use a developed hydrogel framework-assisted digital optofluidic system to capture in situ monitoring of cellular spatial photon dissipation over extended periods. Our findings reveal a dynamic spatial photon dissipation evolution process, presenting a “seesaw” photon dissipation behavior and identifying a regulatory mechanism involving dynamic transitions of photon dissipation patterns for flexible spatial light-energy dissipation management. Additionally, electron transport inhibition and osmotic-stressed experiments demonstrate that this mechanism can be artificially modulated, allowing for control of light-energy dissipation. Our research elucidates the understanding of this vital biological process and defines its features, offering crucial insights into its role in driving entropy production in the biosphere.
The photon dissipation process in photosynthetic reactions has been recognized for its biological significance in contributing to greater biosphere entropy production. While substantial progress has been made in understanding the temporal dynamics of its dissipation regulation, the spatial dynamics remain poorly understood. Here, we use a developed hydrogel framework-assisted digital optofluidic system to capture in situ monitoring of cellular spatial photon dissipation over extended periods. Our findings reveal a dynamic spatial photon dissipation evolution process, presenting a “seesaw” photon dissipation behavior and identifying a regulatory mechanism involving dynamic transitions of photon dissipation patterns for flexible spatial light-energy dissipation management. Additionally, electron transport inhibition and osmotic-stressed experiments demonstrate that this mechanism can be artificially modulated, allowing for control of light-energy dissipation. Our research elucidates the understanding of this vital biological process and defines its features, offering crucial insights into its role in driving entropy production in the biosphere.
关键词：
Machine learning;Transition-metal-dichalcogenide sheets;Density functional theory;Binding energy;Charge transfer
摘要：
With the continuous development of data science, data-driven research methods based on machine learning (ML) have introduced a new paradigm for materials research. Among these, approaches utilizing self-constructed small datasets have gained considerable attention. In this study, we simulated the doping of 28 transition metal atoms on three transition-metal-dichalcogenide (TMD) substrates using Materials Studio software, based on density functional theory (DFT), to generate a small dataset of 168 instances. This dataset was used to explore the predictive capabilities of ML algorithms for binding energy and charge transfer in doped systems. After optimizing hyperparameters through 10-fold cross-validation on the training set, six ML models with better performance were selected for further evaluation on the test set. These models were then applied to predict the binding energy and charge transfer of five transition metals on three new TMD substrates. The results showed that two independently trained Extreme Gradient Boosting (EGB) models performed the best, with MSE, RMSE , MAE, and R² values of 1.056, 1.028, 0.903, and 0.299 for binding energy prediction, and 0.005, 0.085, 0.051, and 0.741 for charge transfer prediction, respectively. The relatively low R² value for binding energy prediction indicates room for improvement, potentially through dataset expansion or additional descriptors, while the higher R² value for charge transfer prediction reflects good model generalization, closely matching the training set performance. Further analysis identified ε d as the most influential descriptor for charge transfer predictions.
With the continuous development of data science, data-driven research methods based on machine learning (ML) have introduced a new paradigm for materials research. Among these, approaches utilizing self-constructed small datasets have gained considerable attention. In this study, we simulated the doping of 28 transition metal atoms on three transition-metal-dichalcogenide (TMD) substrates using Materials Studio software, based on density functional theory (DFT), to generate a small dataset of 168 instances. This dataset was used to explore the predictive capabilities of ML algorithms for binding energy and charge transfer in doped systems. After optimizing hyperparameters through 10-fold cross-validation on the training set, six ML models with better performance were selected for further evaluation on the test set. These models were then applied to predict the binding energy and charge transfer of five transition metals on three new TMD substrates. The results showed that two independently trained Extreme Gradient Boosting (EGB) models performed the best, with MSE, RMSE , MAE, and R² values of 1.056, 1.028, 0.903, and 0.299 for binding energy prediction, and 0.005, 0.085, 0.051, and 0.741 for charge transfer prediction, respectively. The relatively low R² value for binding energy prediction indicates room for improvement, potentially through dataset expansion or additional descriptors, while the higher R² value for charge transfer prediction reflects good model generalization, closely matching the training set performance. Further analysis identified ε d as the most influential descriptor for charge transfer predictions.
关键词：
Dissolved gases in oil;Machine learning;Transition metal dichalcogenides;Gas sensor materials;DFT
摘要：
Rapid advances in 2D materials and elemental doping technologies have developed a wide range of potential sensor materials for detecting gases. However, the challenge is to explore possible material combinations efficiently. While first-principles calculations can accurately predict molecular-scale properties and guide experimental material design, they are too computationally expensive to evaluate all possible transition metal-doped 2D materials. Data-based machine learning methods bring a new approach to this dilemma. In our work, for the exploration of gas-sensitive materials for oil-immersed transformer failures, we constructed datasets of adsorption energies of four transformer dissolved gases on transition-metal dichalcogenides (TMDs) doped with transition metals and trained ten common classification models for adsorption energy prediction. By comparing precision, accuracy, recall, and f1-score, we identified three well-performing models: Bagging, Voting, and RandomForest. The ROC curves and confusion matrix analyses show that the Bagging has the most balanced performance among all prediction categories. The results of the three models analyzed by SHAP show that the newly introduced CM significantly impact the adsorption energy prediction and can be used as a new descriptor for adsorption energy prediction. Finally, we used models to predict the adsorption energies on TM-doped WSe 2 substrates, and a series of analyses identified Au–WSe 2 (C 2 H 2 and C 2 H 4 ), Cu–WSe 2 (CO), and Ni–WSe 2 (H 2 ) as promising sensor materials. Our work provides a powerful tool for the efficient screening of promising materials for dissolved gas sensors in oil, and the application of machine learning in materials science provides experience and methodology.
Rapid advances in 2D materials and elemental doping technologies have developed a wide range of potential sensor materials for detecting gases. However, the challenge is to explore possible material combinations efficiently. While first-principles calculations can accurately predict molecular-scale properties and guide experimental material design, they are too computationally expensive to evaluate all possible transition metal-doped 2D materials. Data-based machine learning methods bring a new approach to this dilemma. In our work, for the exploration of gas-sensitive materials for oil-immersed transformer failures, we constructed datasets of adsorption energies of four transformer dissolved gases on transition-metal dichalcogenides (TMDs) doped with transition metals and trained ten common classification models for adsorption energy prediction. By comparing precision, accuracy, recall, and f1-score, we identified three well-performing models: Bagging, Voting, and RandomForest. The ROC curves and confusion matrix analyses show that the Bagging has the most balanced performance among all prediction categories. The results of the three models analyzed by SHAP show that the newly introduced CM significantly impact the adsorption energy prediction and can be used as a new descriptor for adsorption energy prediction. Finally, we used models to predict the adsorption energies on TM-doped WSe 2 substrates, and a series of analyses identified Au–WSe 2 (C 2 H 2 and C 2 H 4 ), Cu–WSe 2 (CO), and Ni–WSe 2 (H 2 ) as promising sensor materials. Our work provides a powerful tool for the efficient screening of promising materials for dissolved gas sensors in oil, and the application of machine learning in materials science provides experience and methodology.
作者机构：
[Liang Yan; Haiheng Liu] School of Electrical and Electronic Engineering, Wuhan Polytechnic University, Wuhan 430023, China;[Xinyue Liu; Ling Miao; Qiuyun Fu] National Demonstrative School of Miceoelectronics & Wuhan National Laboratory for Optoelectronics & Engineering Research Center for Functional Ceramics of the Ministry of Education, School of Integrated Circuits, Huazhong University of Science and Technology, Wuhan, 430074, China
通讯机构：
[Qiuyun Fu] N;National Demonstrative School of Miceoelectronics & Wuhan National Laboratory for Optoelectronics & Engineering Research Center for Functional Ceramics of the Ministry of Education, School of Integrated Circuits, Huazhong University of Science and Technology, Wuhan, 430074, China
摘要：
As a strongly correlated material, SmNiO 3 has emerged as a promising candidate for non-cryogenic infrared detectors. However, strategies for tuning its properties remain relatively underexplored. In this study, first-principles calculations were employed to investigate the effects of low-concentration Bi doping in SmNiO 3 , revealing several intriguing physical phenomena. Analyses of the density of states, band structure, and charge density confirm that Bi doping introduces localized interband states. These states increase the hole effective mass while reducing the carrier concentration, resulting in an initial rise followed by a subsequent decline in the insulating phase resistivity. Further investigation shows that low-concentration Bi-doping enhances the disparity among different NiO 6 octahedra, thereby increasing the resistance modulation ratio during the metal-insulator phase transition. Additionally, the results indicate a tendency for Bi atoms to cluster on (001) crystallographic planes. These findings provide a novel foundation for advancing SmNiO 3 -based non-cryogenic infrared detectors and related nickelate perovskites.
As a strongly correlated material, SmNiO 3 has emerged as a promising candidate for non-cryogenic infrared detectors. However, strategies for tuning its properties remain relatively underexplored. In this study, first-principles calculations were employed to investigate the effects of low-concentration Bi doping in SmNiO 3 , revealing several intriguing physical phenomena. Analyses of the density of states, band structure, and charge density confirm that Bi doping introduces localized interband states. These states increase the hole effective mass while reducing the carrier concentration, resulting in an initial rise followed by a subsequent decline in the insulating phase resistivity. Further investigation shows that low-concentration Bi-doping enhances the disparity among different NiO 6 octahedra, thereby increasing the resistance modulation ratio during the metal-insulator phase transition. Additionally, the results indicate a tendency for Bi atoms to cluster on (001) crystallographic planes. These findings provide a novel foundation for advancing SmNiO 3 -based non-cryogenic infrared detectors and related nickelate perovskites.
摘要：
Graph clustering is a critical task in network analysis, aimed at grouping nodes based on their structural or attributed similarities. In particular, attributed graph clustering, which considers both structural links and node attributes, is essential for complex networks where additional node information is available. Nonnegative Matrix Factorization (NMF) has shown promise in graph clustering; however, it faces limitations when applied to attributed graph clustering, such as an inability to detect outliers, distortion of geometric data point structures, and disregard for attributed information. Moreover, many existing attributed graph clustering methods overlook distant node relationships due to network sparsity, which hinders further performance improvements. To address these challenges, this paper introduces Weighted Symmetric NMF with graph-Boosting for attributed Graph Clustering (WSBGC), an innovative extension of NMF. WSBGC leverages attribute similarity among nodes to create a weighted version of NMF, enabling outlier detection while preserving the geometric structure of data points. Additionally, WSBGC employs graph-boosting, leveraging attribute information to account for distant node relationships and improve clustering accuracy. A graph attention auto-encoder is then used to construct the final clustering model. The effectiveness of WSBGC is validated through extensive experiments on real-world datasets. Notably, our algorithm improves accuracy by 2.5% compared to the best available method, demonstrating its superior clustering performance in attributed graphs.
Graph clustering is a critical task in network analysis, aimed at grouping nodes based on their structural or attributed similarities. In particular, attributed graph clustering, which considers both structural links and node attributes, is essential for complex networks where additional node information is available. Nonnegative Matrix Factorization (NMF) has shown promise in graph clustering; however, it faces limitations when applied to attributed graph clustering, such as an inability to detect outliers, distortion of geometric data point structures, and disregard for attributed information. Moreover, many existing attributed graph clustering methods overlook distant node relationships due to network sparsity, which hinders further performance improvements. To address these challenges, this paper introduces Weighted Symmetric NMF with graph-Boosting for attributed Graph Clustering (WSBGC), an innovative extension of NMF. WSBGC leverages attribute similarity among nodes to create a weighted version of NMF, enabling outlier detection while preserving the geometric structure of data points. Additionally, WSBGC employs graph-boosting, leveraging attribute information to account for distant node relationships and improve clustering accuracy. A graph attention auto-encoder is then used to construct the final clustering model. The effectiveness of WSBGC is validated through extensive experiments on real-world datasets. Notably, our algorithm improves accuracy by 2.5% compared to the best available method, demonstrating its superior clustering performance in attributed graphs.
摘要：
Doping semiconductor quantum dots (QDs) with heteroatoms has emerged as a technologically important tool for tuning their electronic structure and improving their photocatalytic activities; however, the resulting systems still suffer from low efficiency towards photocatalytic CO 2 reduction. Herein, we report that co-doping CdS QDs with Se 2- and Ga 3+ can boost the photocatalytic CO 2 reduction. The achieved CO evolution rate over Ga/Se co-doped CdS QDs is almost 20 times higher than that of pristine CdS QDs, and is also higher than some other reported QDs based photocatalysts for CO 2 photoreduction. The mechanism study uncovers that the co-doping of Se and Ga in CdS QDs can suppress the recombination of photogenerated electron–hole pairs and decrease the formation barrier of the key *COOH intermediate for promoting the CO 2 to CO conversion.
Doping semiconductor quantum dots (QDs) with heteroatoms has emerged as a technologically important tool for tuning their electronic structure and improving their photocatalytic activities; however, the resulting systems still suffer from low efficiency towards photocatalytic CO 2 reduction. Herein, we report that co-doping CdS QDs with Se 2- and Ga 3+ can boost the photocatalytic CO 2 reduction. The achieved CO evolution rate over Ga/Se co-doped CdS QDs is almost 20 times higher than that of pristine CdS QDs, and is also higher than some other reported QDs based photocatalysts for CO 2 photoreduction. The mechanism study uncovers that the co-doping of Se and Ga in CdS QDs can suppress the recombination of photogenerated electron–hole pairs and decrease the formation barrier of the key *COOH intermediate for promoting the CO 2 to CO conversion.
摘要：
The existing deep learning-based Airborne Laser Scanning (ALS) point cloud semantic segmentation methods require a large amount of labeled data for training, which is not always feasible in practice. Insufficient training data may lead to over-fitting. To address this issue, we propose a novel Multi-branch Feature Extractor (MFE) and a three-stage transfer learning strategy that conducts pre-training on multi-source ALS data and transfers the model to another dataset with few samples, thereby improving the model's generalization ability and reducing the need for manual annotation. The proposed MFE is based on a novel multi-branch architecture integrating Neighborhood Embedding Block (NEB) and Point Transformer Block (PTB); it aims to extract heterogeneous features (e.g., geometric features, reflectance features, and internal structural features) by leveraging the parameters contained in ALS point clouds. To address model transfer, a three-stage strategy was developed: (1) A pre-training subtask was employed to pre-train the proposed MFE if the source domain consisted of multi-source ALS data, overcoming parameter differences. (2) A domain adaptation subtask was employed to align cross-domain feature distributions between source and target domains. (3) An incremental learning subtask was proposed for continuous learning of novel categories in the target domain, avoiding catastrophic forgetting. Experiments conducted on the source domain consisted of DALES and Dublin datasets and the target domain consists of ISPRS benchmark dataset. The experimental results show that the proposed method achieved the highest OA of 85.5% and an average F1 score of 74.0% using only 10% training samples, which means the proposed framework can reduce manual annotation by 90% while keeping competitive classification accuracy.
摘要：
To achieve real-time detection on resource-constrained edge devices, a lightweight vehicle detection algorithm named YOLO-edge was developed based on the YOLOv5s framework. In YOLO-edge, a slim neck was designed to reduce the computational cost in the feature channel fusion process. A modified fast spatial pyramid pooling technique enhances feature extraction efficiency. A multi-scale feature fusion architecture with a broad receptive field integrates shallow and deep features, preserving details for comprehensive feature representation. The loss function was optimized to improve the accuracy of bounding box predictions. Experimental results demonstrate that YOLO-edge achieves a 5.1% increase in mean average precision (mAP) compared to YOLOv5s. Real-time detection rates of 34 and 47 frames per second (FPS) are achieved on edge devices Jetson TX2 and Jetson Orin NX, respectively. YOLO-edge exhibited superior speed and accuracy on edge devices, outperforming state-of-the-art YOLO methods in vehicle detection.
通讯机构：
[Li, SS ] W;Wuhan Polytech Univ, Sch Elect & Elect Engn, Wuhan 430048, Peoples R China.
关键词：
bidirectional feature fusion;defect detection;dynamic histogram attention;low-carbon operation and maintenance;transmission line insulators
摘要：
Against the background of the "dual carbon" goal and the construction of a new power system, the intelligent operation and maintenance of insulators for ultra-high voltage transmission lines face challenges such as difficulty in detecting small-scale defects and strong interference from complex backgrounds. This paper proposes an improved network IDD-DETR to address the problems of inefficient one-way feature fusion and low-contrast defects that are easily overwhelmed in existing RT-DETR models. The enhanced network IDD-DETR replaces PAFPN with a Feature-Focused Diffusion Network (FFDN) and improves multi-scale fusion efficiency through bidirectional cross-scale interaction and designs Dynamic-Range Histogram Self-Attention (DHSA) to enhance defect response in low brightness areas. The experiment showed that its mAP(50) reached 81.7% (an increase of 3.8% percentage points compared to RT-DETR), the flashover defect AP(50) reached 74.6% (+6.1% percentage points), and it maintained 76 FPS on NVIDIA RTX3060, with an average decrease of 1.65% in mAP(50) under complex environments. This model reduces the comprehensive missed detection rate from 26.7% to 23.3%, reduces 45.6 GWh of power loss annually (corresponding to 283,000 tons of CO(2) emission reductions, with 64.3% of the reduction contributed by flashover defect detection), improves inspection efficiency by 60%, reduces manual pole climbing frequency by 37%, and reduces 28 high-altitude risk events annually, providing support for low-carbon operation and maintenance of transmission lines.
作者机构：
[Shaofeng Bian] Key Laboratory of Geological Survey and Evaluation of Ministry of Education, China University of Geosciences, Wuhan, China;School of Geography and Information Engineering, China University of Geosciences, Wuhan, China;[Yi Liu] School of Electrical and Electronic Engineering, Wuhan Polytechnic University, Wuhan, China;[Yongbing Chen] Department of Navigation Engineering, Naval University of Engineering, Wuhan, China;[Zemin Wu] Key Laboratory of Geological Survey and Evaluation of Ministry of Education, China University of Geosciences, Wuhan, China<&wdkj&>School of Geography and Information Engineering, China University of Geosciences, Wuhan, China
通讯机构：
[Shaofeng Bian] K;Key Laboratory of Geological Survey and Evaluation of Ministry of Education, China University of Geosciences, Wuhan, China
摘要：
Integer ambiguity resolution plays a key role in applications related to GNSS precise positioning. This contribution focuses on three commonly used integer estimators (IEs), i.e., integer rounding (IR), integer bootstrapping (IB) and integer least-squares (ILS). Contributions are mainly of four aspects. First, the objective function of IR and IB is given, respectively. Second, an upper bound for ILS is proposed. Third, a sorting technique is introduced to tighten the upper bound of IR/ILS after decorrelation. Fourth, the success-rate approximation for IR and ILS with bounded error is developed, respectively. Finally, real-collected data in PPP validate the following arguments. (i) Applying sorting technique after decorrelation can improve the tightness of IR upper bound a lot, but can only slightly tighten the ILS upper bound. (ii) The proposed ILS upper bound is almost as the same tight as the one in PS-LAMBDA software, much tighter than other known upper bounds. Unlike the PS-LAMBDA upper bound, the proposed ILS upper bound has the advantage of time efficiency for real-time applications. (iii) The proposed approximations produce more accuracy approximated success-rate than the frequently-used ambiguity dilution of precision (ADOP) based approximation.
期刊：
Frontiers in Artificial Intelligence,2025年8:1463233 ISSN：2624-8212
通讯作者：
Cao, L
作者机构：
[Li, Wei; Cao, Li] School of Electrical and Electronic Engineering, Wuhan Polytechnic University, Wuhan, China;[Deng, He] School of Computer Science and Technology, Wuhan University of Science and Technology, Wuhan, China
通讯机构：
[Cao, L ] W;Wuhan Polytech Univ, Sch Elect & Elect Engn, Wuhan, Peoples R China.
关键词：
U-Net;feature-location attention;foveal avascular zone (FAZ) segmentation;joint loss function;optical coherence tomography angiography (OCTA)
摘要：
INTRODUCTION: Since optical coherence tomography angiography (OCTA) is non-invasive and non-contact, it is widely used in the study of retinal disease detection. As a key indicator for retinal disease detection, accurate segmentation of foveal avascular zone (FAZ) has an important impact on clinical application. Although the U-Net and its existing improvement methods have achieved good performance on FAZ segmentation, their generalization ability and segmentation accuracy can be further improved by exploring more effective improvement strategies. METHODS: We propose a novel improved method named Feature-location Attention U-Net (FLA-UNet) by introducing new designed feature-location attention blocks (FLABs) into U-Net and using a joint loss function. The FLAB consists of feature-aware blocks and location-aware blocks in parallel, and is embed into each decoder of U-Net to integrate more marginal information of FAZ and strengthen the connection between target region and boundary information. The joint loss function is composed of the cross-entropy loss (CE loss) function and the Dice coefficient loss (Dice loss) function, and by adjusting the weights of them, the performance of the network on boundary and internal segmentation can be comprehensively considered to improve its accuracy and robustness for FAZ segmentation. RESULTS: The qualitative and quantitative comparative experiments on the three datasets of OCTAGON, FAZID and OCTA-500 show that, our proposed FLA-UNet achieves better segmentation quality, and is superior to other existing state-of-the-art methods in terms of the MIoU, ACC and Dice coefficient. DISCUSSION: The proposed FLA-UNet can effectively improve the accuracy and robustness of FAZ segmentation in OCTA images by introducing feature-location attention blocks into U-Net and using a joint loss function. This has laid a solid theoretical foundation for its application in auxiliary diagnosis of fundus diseases.
作者机构：
[Yage Qie; Jinghua Huang; Donghao Wu; Jian Jiang] School of Electrical and Electronic Engineering, Wuhan Polytechnic University, Wuhan 430024, China;[Chao Fang] Author to whom correspondence should be addressed.
通讯机构：
[Chao Fang] A;Author to whom correspondence should be addressed.
摘要：
The field of underwater robotics is experiencing rapid growth, wherein accurate object detection constitutes a fundamental component. Given the prevalence of false alarms and omission errors caused by intricate subaquatic conditions and substantial image noise, this study introduces an enhanced detection framework that combines the YOLOv8 architecture with multimodal visual fusion methodology. To solve the problem of degraded detection performance of the model in complex environments like those with low illumination, features from Visible Light Image are fused with the Thermal Distribution Features exhibited by Infrared Image, thereby yielding more comprehensive image information. Furthermore, to precisely focus on crucial target regions and information, a Multi-Scale Cross-Axis Attention Mechanism (MSCA) is introduced, which significantly enhances Detection Accuracy. Finally, to meet the lightweight requirement of the model, an Efficient Shared Convolution Head (ESC_Head) is designed. The experimental findings reveal that the YOLOv8-FUSED framework attains a mean average precision (mAP) of 82.1%, marking an 8.7% enhancement compared to the baseline YOLOv8 architecture. The proposed approach also exhibits superior detection capabilities relative to existing techniques while simultaneously satisfying the critical requirement for real-time underwater object detection. Moreover, the proposed system successfully meets the essential criteria for real-time detection of underwater objects.
作者机构：
[Zhao, YuDan; Ni, Ying; Xia, Peng] Wuhan Polytech Univ, Sch Elect & Elect Engn, Wuhan 430023, Hubei, Peoples R China.;[Zeng, Wu; Zeng, W] Wuhan Polytech Univ, Sch Math & Comp Sci, Wuhan 430023, Hubei, Peoples R China.;[Tan, RuoChen] Univ Calif San Diego, Comp Sci & Engn, San Diego, CA 92093 USA.
会议名称：
1st International Artificial Intelligence Conference-IAIC
会议时间：
NOV 25-27, 2023
会议地点：
Nanjing, PEOPLES R CHINA
会议主办单位：
[Zhao, YuDan;Ni, Ying;Xia, Peng] Wuhan Polytech Univ, Sch Elect & Elect Engn, Wuhan 430023, Hubei, Peoples R China.^[Zeng, Wu] Wuhan Polytech Univ, Sch Math & Comp Sci, Wuhan 430023, Hubei, Peoples R China.^[Tan, RuoChen] Univ Calif San Diego, Comp Sci & Engn, San Diego, CA 92093 USA.
会议论文集名称：
Communications in Computer and Information Science
关键词：
Data Visualization;Hydrological Data;Digital Twin
摘要：
In the context of digital transformation, cities and enterprises are striving to build a digital industrial chain, cultivate a digital ecosystem, and support high-quality economic development. Therefore, the use of visualization technology to assist decision-makers in rational planning has become a hot spot. Taking wuhan city as an example, combined with 3D modeling technology, it is aimed at smart cities and based on digital twins to create multiple scenarios for hydrological data application services and improve hydrological information services. First, we collected the data released by the china hydrology and water resources station; then, we visualized the hydrological data of the yangtze river Hankou station by using methods such as view juxtaposition and 3D interaction; after that, we constructed a 3D scene based on the real scene of the yangtze river Hankou basin, and used algorithms to the water body model is optimized; finally, the interaction between data and scenes is designed, various functions are realized by using high-level programming language design, and the water level changes in the flood season are simulated to help analyze and understand data more clearly, and assist decision makers in making decisions.
关键词：
metasurface;deep neural network;acoustic field modulation;inverse design;genetic algorithm
摘要：
Existing research in metasurface design was based on trial-and-error high-intensity iterations and requires deep acoustic expertise from the researcher, which severely hampered the development of the metasurface field. Using deep learning enabled the fast and accurate design of hypersurfaces. Based on this, in this paper, an integrated learning approach was first utilized to construct a model of the forward mapping relationship between the hypersurface physical structure parameters and the acoustic field, which was intended to be used for data enhancement. Then a dual-feature fusion model (DFCNN) based on a convolutional neural network was proposed, in which the first feature was the high-dimensional nonlinear features extracted using a data-driven approach, and the second feature was the physical feature information of the acoustic field mined using the model. A convolutional neural network was used for feature fusion. A genetic algorithm was used for network parameter optimization. Finally, generalization ability verification was performed to prove the validity of the network model. The results showed that 90% of the integrated learning models had an error of less than 3 dB between the real and predicted sound field data, and 93% of the DFCNN models could achieve an error of less than 5 dB in the local sound field intensity.
摘要：
Few-layer tin (Sn)-based nanosheets (NSs) with a thickness of approximate to 2.5 nm are successfully prepared using a modified liquid phase exfoliation (LPE) method. Here the first exploration of photo-electrochemical (PEC) and nonlinear properties of Sn NSs is presented. The results demonstrate that the PEC properties are tunable under different experimental conditions. Additionally, Sn NSs are shown to exhibit a unique self-powered PEC performance, maintaining a good long-term stability for up to 1 month. Using electron spin resonance, active species, such as hydroxyl radicals (<middle dot>OH), superoxide radicals (<middle dot>O2-), and holes (h+), are detected during operations, providing a deeper understanding of the working mechanism. Furthermore, measurements of nonlinear response reveal that Sn NSs can be effective for all-optical modulation, as it enables the realization of all-optical switching through excitation spatial cross-phase modulation (SXPM). These findings present new research insights and potential applications of Sn NSs in optoelectronics. 2D tin nanosheets (Sn NSs) with a thickness of around 2.5 nm is prepared and applied for photo-electrochemical (PEC) and all-optical modulation applications for the first time. Sn NSs show unique self-powered PEC photodetection performance, and a working mechanism is proposed. Sn NSs also show attractive nonlinear properties, highlighting its potential for designing all-optical switches. image
通讯机构：
[Ma, HC ] W;Wuhan Univ, Sch Remote Sensing & Informat Engn, Wuhan 430079, Peoples R China.
关键词：
Buildings;Extraction;Shape optimization;Topology;Alpha-shapes;Boundary point extraction;Boundary-points;Contextual topological optimization;Direction detections;Dominant direction detection;LiDAR;Outline extractions;Point extraction;Topological optimization;Optical radar
摘要：
It is challenging to extract satisfactory building outlines from LiDAR data due to the unorganized point cloud and complex building shapes. To solve the issues, a method using adaptive tracing alpha shapes (ATAS) and contextual topological optimization is proposed. First, the ATAS method is used to extract sequential boundary points. After that, a method based on point cloud distribution analysis is developed to obtain building dominant directions and line segments of outlines. Finally, regularized outlines are obtained by adjusting all line segments simultaneously under the framework of global energy optimization that considers the geometric errors and contextual geometric relationships between adjacent line segments. Experimental results verify that the proposed ATAS method can efficiently extract sequential boundary points with a minimum 98.49% correctness. In addition, the extracted outlines are attractive and the minimum values of the RMSE , PoLiS, and RCC metrics of the extracted outlines are 0.48 m, 0.44 m, and 0.31 m, respectively, showing the effectiveness of the proposed method.
It is challenging to extract satisfactory building outlines from LiDAR data due to the unorganized point cloud and complex building shapes. To solve the issues, a method using adaptive tracing alpha shapes (ATAS) and contextual topological optimization is proposed. First, the ATAS method is used to extract sequential boundary points. After that, a method based on point cloud distribution analysis is developed to obtain building dominant directions and line segments of outlines. Finally, regularized outlines are obtained by adjusting all line segments simultaneously under the framework of global energy optimization that considers the geometric errors and contextual geometric relationships between adjacent line segments. Experimental results verify that the proposed ATAS method can efficiently extract sequential boundary points with a minimum 98.49% correctness. In addition, the extracted outlines are attractive and the minimum values of the RMSE , PoLiS, and RCC metrics of the extracted outlines are 0.48 m, 0.44 m, and 0.31 m, respectively, showing the effectiveness of the proposed method.
摘要：
In the last few decades, nanoparticles have been a prominent topic in various fields, particularly in agriculture, due to their unique physicochemical properties. Herein, molybdenum copper lindgrenite Cu(3)(MoO(4))(2)(OH)(2) (CM) nanoflakes (NFs) are synthesized by a one-step reaction involving α-MoO(3) and CuCO(3)⋅Cu(OH)(2)⋅xH(2)O solution at low temperature for large scale industrial production and developed as an effective antifungal agent for the oilseed rape. This synthetic method demonstrates great potential for industrial applications. Infrared spectroscopy and X-ray diffraction (XRD) results reveal that CM samples exhibit a pure monoclinic structure. TG and DSC results show the thermal stable properties. It can undergo a phase transition form copper molybdate (Cu(3)Mo(2)O(9)) at about 300°C. Then Cu(3)Mo(2)O(9) nanoparticlesdecompose into at CuO and MoO(3) at 791°C. The morphology of CM powder is mainly composed of uniformly distributed parallelogram-shaped nanoflakes with an average thickness of about 30nm. Moreover, the binding energy of CM NFs is measured to be 2.8eV. To assess the antifungal properties of these materials, both laboratory and outdoor experiments are conducted. In the pour plate test, the minimum inhibitory concentration (MIC) of CM NFs against Sclerotinia sclerotiorum (S. sclerotiorum) is determined to be 100ppm, and the zone of inhibiting S. sclerotiorum is 14mm. When the concentration is above 100nm, the change rate of the hyphae circle slows down a little and begins to decrease until to 200ppm. According to the aforementioned findings, the antifungal effects of a nano CM NFs solution are assessed at different concentrations (0ppm (clear water), 40ppm, and 80ppm) on the growth of oilseed rape in an outdoor setting. The results indicate that the application of CM NFs led to significant inhibition of S. sclerotiorum. Specifically, when the nano CM solution was sprayed once at the initial flowering stage at a concentration of 80ppm, S. sclerotiorum growth was inhibited by approximately 34%. Similarly, when the solution was sprayed once at the initial flowering stage and once at the rape pod stage, using a concentration of 40ppm, a similar level of inhibition was achieved. These outcomes show that CM NFs possess the ability to bind with more metal ions due to their larger specific surface area. Additionally, their semiconductor physical properties enable the generation of reactive oxygen species (ROS). Therefore, CM NFs hold great potential for widespread application in antifungal products.
期刊：
E3S Web of Conferences,2024年520:02022 ISSN：2555-0403
作者机构：
[Guoqiang Hao; Zhen Huang; Wei Chen; Qiang Lv] School of Electrical and Electronic Engineering, Wuhan Polytechnic University;[Feng Zheng] KINGDREAM PUBLIC LIMITED COMPANY
会议名称：
第四届环境资源与能源工程国际学术会议
会议时间：
2024-02-23
会议地点：
中国广东珠海
关键词：
Stacking;domain;Estimation of Stacked Fusion Model of Young's Modulus;Measurement of Rock Deformation Parameters;Fusion;Model
摘要：
Rock Young’s modulus is an essential parameter for formation stress characterization and oil and gas reservoir evaluation work and plays an important role in oil drilling-related engineering type work. Aiming at the problems of doubtful confidence in Young’s modulus measurements, time-consuming computation, and high measurement cost in oil drilling, this paper proposed Young’s modulus estimation method based on the Stacking fusion model. The method first processed the downhole vibration data to obtain its time-domain feature data and then used the time-domain feature data as the input to the fusion model while used the rock Young’s modulus data as the model output. The model learner used consists of three base learners, ANN, XGBoost, and CatBoost, with MLR as the model meta-learner. The mapping relationship between the time-domain features and Young’s modulus was established by this method, and the prediction and estimation of Young’s modulus parameters of the rock were finally realized. The results showed that the average absolute error (MAE) of the fused Stacking model was 0.2502 and the goodness-of-fit (R2) was 0.9691. Compared with other single models, the fused model based on Stacking had the advantage of being able to combine each single model, which provided a new method for estimation and prediction of Young’s modulus of rocks.
