作者机构:
[Lin Jia; Wenjun Jia; Wuao Tang] School of Navigation, Wuhan University of Technology, Wuhan 430063, China;Hubei Key Laboratory of Inland Shipping Technology, Wuhan 430063, China;School of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430048, China;[Lichuan Wu] Department of Earth Sciences, Uppsala University, Uppsala 75236, Sweden;[Chunhui Zhou] School of Navigation, Wuhan University of Technology, Wuhan 430063, China<&wdkj&>Hubei Key Laboratory of Inland Shipping Technology, Wuhan 430063, China
通讯机构:
[Hongxun Huang] H;Hubei Key Laboratory of Inland Shipping Technology, Wuhan 430063, China<&wdkj&>School of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430048, China
关键词:
To visualize the research hotspots and development direction of green ships in the fields of technology;operation and policy measures;this subsection applies the Co-occurrence Analysis of VOSviewer to generate a visual map. VOSviewer software's clustering algorithm is primarily based on the correlation intensity algorithm;which selects high-frequency keywords in the text for clustering analysis;and reflects the research theme of the discipline through high-frequency keywords. The analysis
摘要:
With the increased global focus on environmental protection, green ships have emerged as the cornerstone to the maritime industry's long-term development. Based on the knowledge graph method, this paper uses bibliometric analysis to examine the research progress of green ships in a multidimensional perspective. Firstly, sort out the definition and connotation of green ships, and clarify their greening requirements in the whole life cycle. Bibliometric analysis of relevant literature using the bibliometrix package; secondly, combined with the virtual analysis method-VOSviewer software, it explores the current research status of green ships from the multi-dimensional perspectives of technical, operational and policy. The results indicate that the hot spots in technical measures are “green energy”, “ship optimization” and “waste heat recovery”; the hot spots in operational measures are “speed optimization” and “liner shipping”; and the hot spots in policy measures are “market mechanism” and “international regulations”.
With the increased global focus on environmental protection, green ships have emerged as the cornerstone to the maritime industry's long-term development. Based on the knowledge graph method, this paper uses bibliometric analysis to examine the research progress of green ships in a multidimensional perspective. Firstly, sort out the definition and connotation of green ships, and clarify their greening requirements in the whole life cycle. Bibliometric analysis of relevant literature using the bibliometrix package; secondly, combined with the virtual analysis method-VOSviewer software, it explores the current research status of green ships from the multi-dimensional perspectives of technical, operational and policy. The results indicate that the hot spots in technical measures are “green energy”, “ship optimization” and “waste heat recovery”; the hot spots in operational measures are “speed optimization” and “liner shipping”; and the hot spots in policy measures are “market mechanism” and “international regulations”.
期刊:
International Journal of Geographical Information Science,2025年 ISSN:1365-8816
通讯作者:
Yao Yao
作者机构:
[Mi Tang; Yueheng Ma; Zhihui Hu] School of Geography and Information Engineering, China University of Geosciences, Wuhan, Hubei Province, China;LocationMind Institution, LocationMind Inc., Chiyoda, Tokyo, Japan;[Huanjun Hu] School of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan, China;National Engineering Research Center of Geographic Information System, China University of Geosciences, Wuhan, Hubei province, China;Hitotsubashi Institute for Advanced Study, Hitotsubashi University, Kunitachi, Tokyo, Japan
通讯机构:
[Yao Yao] S;School of Geography and Information Engineering, China University of Geosciences, Wuhan, Hubei Province, China<&wdkj&>LocationMind Institution, LocationMind Inc., Chiyoda, Tokyo, Japan<&wdkj&>National Engineering Research Center of Geographic Information System, China University of Geosciences, Wuhan, Hubei province, China<&wdkj&>Hitotsubashi Institute for Advanced Study, Hitotsubashi University, Kunitachi, Tokyo, Japan<&wdkj&>Faculty of Engineering, Reitaku University, Kashiwa, Chiba, Japan<&wdkj&>Ministry of Land and Resources of China, Key Laboratory of Urban Land Resources Monitoring and Simulation, Shenzhen, China
关键词:
Land-use classification;LandGPT;LLM tuning;prompt engineering;multi-source geospatial data
摘要:
Actual land parcels vary significantly in size and complexity. Previous studies were limited by existing technical methods for fine-grained land use classification. The emergence of multimodal large language models offers new techniques for image classification, but their application in land use classification remains unexplored. This study presents LandGPT, a multimodal large language model trained on the CN-MSLU-100K dataset, covering fine-grained land use classification of irregular parcels. This study proposes a trans-level discrimination framework to improve LandGPT’s ability to classify fine-grained land use. Under this framework, LandGPT achieves a discrimination accuracy of 89.7% and a Kappa coefficient of 0.85 for fine-grained land use categories, outperforming state-of-the-art models with a 48.33% accuracy improvement. In some challenging categories, the improvement reaches nearly 1500%. This study finds that training with multi-source remote sensing image data improved LandGPT’s accuracy by 15.79% compared to single-image data. This study explores Prompt engineering based on LandGPT. The optimal prompt paradigm offers fine-grained categories and guides the model for accurate classification, reducing errors from LLM hallucinations. This study pioneeringly explores the application of large language models in the land use domain and offers a new solution for fine-grained land use classification.
期刊:
Mechanical Systems and Signal Processing,2025年234:112826 ISSN:0888-3270
通讯作者:
Shan Zeng
作者机构:
[Site Lv; Shan Zeng; Chen Yu; Ke Yang] School of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan, Hubei 430023, China;[Hongan Wu] Hubei Key Laboratory for Efficient Utilization and Agglomeration of Metallurgic Mineral Resources, Wuhan University of Science and Technology, Wuhan 430081, China
通讯机构:
[Shan Zeng] S;School of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan, Hubei 430023, China
摘要:
When mechanical equipment fails, the fault characteristics are often interfered by adjacent components. Therefore, how to well characterize the time-varying laws of multi-component signals containing adjacent components has always been a difficulty and research hotspot in the application of time–frequency analysis (TFA) technologies in mechanical fault diagnosis. In this paper, a new TFA method is proposed, called the Multi-scale chirplet synchroextracting transform (MCSET). On the basis of chirplet transform (CT), by using two additional parameters, the window rotation step is divided into two sections within each window length to more accurately match the instantaneous frequency (IF) trajectory of the nonlinear frequency modulation signal. In this way, the energy-concentrated time–frequency (TF) distribution of the multi-component signal containing adjacent components can be obtained. Moreover, to further improve the TF resolution, a new frequency estimation operator is constructed using the idea of synchronous extraction to more accurately capture the IF variation law of multi-component signals and keep the energy highly concentrated. MCSET can well characterize the dynamic characteristics of components adjacent to the IF trajectory, and as a parameterized TFA technique, it maintains good noise robustness. In simulation and experiments, compared with other similar advanced TFA techniques, the results can verify the effectiveness of the proposed method and its superiority in processing complex non-stationary signals with adjacent IF trajectories.
When mechanical equipment fails, the fault characteristics are often interfered by adjacent components. Therefore, how to well characterize the time-varying laws of multi-component signals containing adjacent components has always been a difficulty and research hotspot in the application of time–frequency analysis (TFA) technologies in mechanical fault diagnosis. In this paper, a new TFA method is proposed, called the Multi-scale chirplet synchroextracting transform (MCSET). On the basis of chirplet transform (CT), by using two additional parameters, the window rotation step is divided into two sections within each window length to more accurately match the instantaneous frequency (IF) trajectory of the nonlinear frequency modulation signal. In this way, the energy-concentrated time–frequency (TF) distribution of the multi-component signal containing adjacent components can be obtained. Moreover, to further improve the TF resolution, a new frequency estimation operator is constructed using the idea of synchronous extraction to more accurately capture the IF variation law of multi-component signals and keep the energy highly concentrated. MCSET can well characterize the dynamic characteristics of components adjacent to the IF trajectory, and as a parameterized TFA technique, it maintains good noise robustness. In simulation and experiments, compared with other similar advanced TFA techniques, the results can verify the effectiveness of the proposed method and its superiority in processing complex non-stationary signals with adjacent IF trajectories.
作者机构:
["Chen, Songnan] School of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan, Hubei, China;["Chen, Songnan] Foshan Zhongke Innovation Research Institute of Intelligent Agriculture and Robotics, Foshan, Guangdong, China;[Feng, Zhaoxu"] China United Network Communications Co., Ltd. Henan Branch, Zhengzhou, Henan, China. iefengzhaoxu@163.com
摘要:
Recent data-driven deep learning methods for image reflection removal have made impressive progress, promoting the quality of photo capturing and scene understanding. Due to the massive consumption of computational complexity and memory usage, the performance of these methods degrades significantly while dealing with high-resolution images. Besides, most existing methods for reflection removal can only remove reflection patterns by downsampling the input image into a much lower resolution, resulting in the loss of plentiful information. In this paper, we propose a novel transformer-based framework for high-resolution image reflection removal, termed as the Laplacian pyramid-based component-aware transformer (LapCAT). LapCAT leverages a Laplacian pyramid network to remove high-frequency reflection patterns and reconstruct the high-resolution background image guided by the clean low-frequency background components. Guided by the reflection mask through pixel-wise contrastive learning, LapCAT designs a component-separable transformer block (CSTB) which removes reflection patterns from the background constituents through a reflection-aware multi-head self-attention mechanism. Extensive experiments on several benchmark datasets for reflection removal demonstrate the superiority of our LapCAT, especially the excellent performance and high efficiency in removing reflection from high-resolution images than state-of-the-art methods.
摘要:
Blended vegetable oil is considered to be a valuable product in the market owing to favourable taste and nutritional composition. The quantification of its contents has notable implications for protecting food safety and consumer interests. Thus, a rapid and non-destructive method is needed to analyse the composition of blended oil. This study established an analytical method combining Raman spectroscopy and prediction models to determine the content of olive oil in a mixture. Competitive adaptive reweighted sampling was employed to select feature bands attributed to β-carotene and unsaturated fatty acids. Various models were used to calculate the mixture proportion, and the importance of characteristic peak intensity affecting the prediction was evaluated via grey relational analysis. The random forest model exhibited superior performance in quantitative analysis, with RMSE and R 2 of 0.0447 and 0.9799, respectively. Overall, this approach was proven to effectively identify blended olive oils, exemplifying its potential in food authentication.
Blended vegetable oil is considered to be a valuable product in the market owing to favourable taste and nutritional composition. The quantification of its contents has notable implications for protecting food safety and consumer interests. Thus, a rapid and non-destructive method is needed to analyse the composition of blended oil. This study established an analytical method combining Raman spectroscopy and prediction models to determine the content of olive oil in a mixture. Competitive adaptive reweighted sampling was employed to select feature bands attributed to β-carotene and unsaturated fatty acids. Various models were used to calculate the mixture proportion, and the importance of characteristic peak intensity affecting the prediction was evaluated via grey relational analysis. The random forest model exhibited superior performance in quantitative analysis, with RMSE and R 2 of 0.0447 and 0.9799, respectively. Overall, this approach was proven to effectively identify blended olive oils, exemplifying its potential in food authentication.
摘要:
In the event of disaster-related disruptions to public networks, an air-ground communication-navigation-sensing (CNS) ad-hoc network (ANET) can effectively support emergency response operations by ensuring communication, navigation and sensing capabilities. However, during extreme events such as typhoons and torrential rain, flight operations become impossible for aircraft. Therefore, a critical issue that requires immediate investigation is how to effectively provide communication, navigation, and sensing assistance to emergency regions using only ground-based CNS nodes. This paper presents a communication-navigation-sensing spatiotemporal cooperative optimization algorithm based on ResNet-DDPG. It aims to achieve spatiotemporal cooperative optimization of ground CNS nodes in the absence of aerial relay ad hoc network nodes (RANET nodes). This work has three main components: (1) a neural network-based geospatial fitting model for ANET communication, (2) a ResNet based spatiotemporal feature and CNS node fusion encoding method for state space construction for reinforcement learning, and (3) a deep reinforcement learning algorithm called ResNet-DDPG for deploying ground ANET nodes dynamically. Experimental results indicate that, compared to similar approaches, this method significantly improves communication rate optimization and deployment site effectiveness. This study supports emergency response by utilizing ground ANET, which is crucial for disaster emergency response, command scheduling, and life rescue in extreme disaster scenarios.
通讯机构:
[Yang, H ] W;Wuhan Polytech Univ, Coll Math & Comp Sci, Wuhan 430023, Peoples R China.
关键词:
BP neural network;hybrid optimization model;metaheuristic algorithm;renewable energy integration;wind power prediction
摘要:
The integration of intermittent wind power into modern grids necessitates highly accurate forecasting models to ensure stability and efficiency. To address the limitations of traditional backpropagation (BP) neural networks, such as slow convergence and susceptibility to local optima, this study proposes a novel hybrid framework: the Multi-Strategy Coati Optimization Algorithm (SZCOA)-optimized BP neural network (SZCOA-BP). The SZCOA integrates three innovative strategies-a population position update mechanism for global exploration, an olfactory tracing strategy to evade local optima, and a soft frost search strategy for refined exploitation-to enhance the optimization efficiency and robustness of BP networks. Evaluated on the CEC2017 benchmark, the SZCOA outperformed state-of-the-art algorithms, including ICOA, DBO, and PSO, achieving superior convergence speed and solution accuracy. Applied to a real-world wind power dataset (912 samples from Alibaba Cloud Tianchi), the SZCOA-BP model attained an R² of 94.437% and reduced the MAE to 10.948, significantly surpassing the standard BP model (R²: 81.167%, MAE: 18.891). Comparative analyses with COA-BP, BWO-BP, and other hybrid models further validated its dominance in prediction accuracy and stability. The proposed framework not only advances wind power forecasting but also offers a scalable solution for optimizing complex renewable energy systems, supporting global efforts toward sustainable energy transitions.
摘要:
The corrosion of steel reinforcement in concrete structures due to chloride exposure poses significant financial and environmental risks. An accurate assessment of chloride diffusion is essential for predicting the service life of steel-reinforced concrete. This article develops six models for estimating the chloride diffusion coefficient (CDC) in concrete, considering various exposure conditions like tidal, splash, atmospheric, and submerged environments. The models utilize a hybrid approach combining the adaptive neuro-fuzzy inference system (ANFIS) and least square support vector regression (LSSVR), enhanced by an improved arithmetic optimization algorithm (IAA). The IAA merges the arithmetic optimization algorithm (AOA) with the Aquila optimizer (AO) to address AOA's limitations. The models undergo sensitivity analysis using the Fourier Amplitude Sensitivity Test (FAST), revealing that the curing mechanism (CM) is the most influential factor, with a sensitivity value of 0.993. The study demonstrates that LSSVR and ANFIS-based methods are highly effective in predicting CDC. Among them, the ANFIS combined with IAA outperforms others regarding reliability and accuracy, making it a superior choice for CDC estimation. This robust model could lead to better management and longevity of concrete structures exposed to chloride, mitigating potential risks. The main application of this research is to enhance the durability management and service life prediction of steel-reinforced concrete structures exposed to chloride-laden environments. Engineers and infrastructure managers can better assess the risk and timing of corrosion onset by accurately estimating the CDC, which is crucial for preventive maintenance and cost-effective design.
摘要:
The accurate detection of road defects is crucial for enhancing the safety and efficiency of road maintenance. This study focuses on six common types of pavement defects: transverse cracks, longitudinal cracks, alligator cracking, oblique cracks, potholes, and repair marks. In real-world scenarios, key challenges include effectively distinguishing between the foreground and background, as well as accurately identifying small-sized (e.g., fine cracks, dense alligator cracking, and clustered potholes) and overlapping defects (e.g., intersecting cracks or clustered damage areas where multiple defects appear close together). To address these issues, this paper proposes a Pavement-DETR model based on the Real-Time Detection Transformer (RT-DETR), aiming to optimize the overall accuracy of defect detection. To achieve this goal, three main improvements are proposed: (1) the introduction of the Channel-Spatial Shuffle (CSS) attention mechanism in the third (S3) and fourth (S4) stages of the ResNet backbone, which correspond to mid-level and high-level feature layers, enabling the model to focus more precisely on road defect features; (2) the adoption of the Conv3XC structure for feature fusion enhances the model's ability to differentiate between the foreground and background, which is achieved through multi-level convolutions, channel expansion, and skip connections, which also contribute to improved gradient flow and training stability; (3) the proposal of a loss function combining Powerful-IoU v2 (PIoU v2) and Normalized Wasserstein Distance (NWD) weighted averaging, where PIoU v2 focuses on optimizing overlapping regions, and NWD targets small object optimization. The combined loss function enables comprehensive optimization of the bounding boxes, improving the model's accuracy and convergence speed. Experimental results show that on the UAV-PDD2023 dataset, Pavement-DETR improves the mean average precision (mAP) by 7.7% at IoU = 0.5, increases mAP by 8.9% at IoU = 0.5-0.95, and improves F1 Score by 7%. These results demonstrate that Pavement-DETR exhibits better performance in road defect detection, making it highly significant for road maintenance work.
作者机构:
[Gao, Rui; Wu, Weixiong; Wu, Peng] Guizhou Beipanjiang Elect Power Co Ltd, Mamaya Photovolta Branch, Guiyang 550081, Peoples R China.;[Yuan, Chen; Xia, Xiaoling] Guizhou New Meteorol Technol Co Ltd, Guiyang 550081, Peoples R China.;[Wang, Yifei; Liu, Renfeng] Wuhan Polytech Univ, Sch Math & Comp Sci, Wuhan 430023, Peoples R China.
通讯机构:
[Yuan, C ] G;Guizhou New Meteorol Technol Co Ltd, Guiyang 550081, Peoples R China.
关键词:
ultra-short-term PV power forecasting;sky imagers;MCR;cloud cover estimation
摘要:
Accurate photovoltaic (PV) power forecasting is crucial for stable grid integration, particularly under rapidly changing weather conditions. This study presents an ultra-short-term forecasting model that integrates sky imager data and meteorological radar data, achieving significant improvements in forecasting accuracy. By dynamically tracking cloud movement and estimating cloud coverage, the model enhances performance under both clear and cloudy conditions. Over an 8-day evaluation period, the average forecasting accuracy improved from 67.26% to 77.47% (+15%), with MSE reduced by 39.2% (from 481.5 to 292.6), R2 increased from 0.724 to 0.855, NSE improved from 0.725 to 0.851, and Theil’s U reduced from 0.42 to 0.32. Notable improvements were observed during abrupt weather transitions, particularly on 1 July and 3 July, where the combination of MCR and sky imager data demonstrated superior adaptability. This integrated approach provides a robust foundation for advancing ultra-short-term PV power forecasting.
摘要:
The visualization of grain components through non-destructive detection is crucial for crop improvement and quality assessment, making it a research focus. Raman chemical imaging, an effective spectral imaging technology, has been extensively applied to detect various grain components. However, practical applications face challenges such as unclear Raman shift-composition relationships, low-quality images affected by noise, and long imaging times. To address these issues, this paper proposes a novel method, Raman Denoising Diffusion Generative Adversarial Network (RDDGAN), based on the Denoising Diffusion Probabilistic Model (DDPM). DDPM is used as the generator, paired with a discriminator to distinguishes generated and real data. This approach rapidly produces high-quality Raman images, highlighting spatial distribution of prominent components and clarifying Raman shift-composition correspondence. Further, this paper takes maize as the research object to carry out experiments. The main components of maize, such as starch, are continuous in spatial distribution, mainly concentrated in endosperm, and have obvious boundary. There are obvious differences in the distribution of embryo and endosperm, thus Raman imaging can be used to effectively characterize maize components. Experiments with maize show that, compared to other algorithms, RDDGAN generates higher-quality images, accurately characterizing the spatial distribution of maize components, which providing a strong foundation for subsequent quality detection and variety improvement applications.
The visualization of grain components through non-destructive detection is crucial for crop improvement and quality assessment, making it a research focus. Raman chemical imaging, an effective spectral imaging technology, has been extensively applied to detect various grain components. However, practical applications face challenges such as unclear Raman shift-composition relationships, low-quality images affected by noise, and long imaging times. To address these issues, this paper proposes a novel method, Raman Denoising Diffusion Generative Adversarial Network (RDDGAN), based on the Denoising Diffusion Probabilistic Model (DDPM). DDPM is used as the generator, paired with a discriminator to distinguishes generated and real data. This approach rapidly produces high-quality Raman images, highlighting spatial distribution of prominent components and clarifying Raman shift-composition correspondence. Further, this paper takes maize as the research object to carry out experiments. The main components of maize, such as starch, are continuous in spatial distribution, mainly concentrated in endosperm, and have obvious boundary. There are obvious differences in the distribution of embryo and endosperm, thus Raman imaging can be used to effectively characterize maize components. Experiments with maize show that, compared to other algorithms, RDDGAN generates higher-quality images, accurately characterizing the spatial distribution of maize components, which providing a strong foundation for subsequent quality detection and variety improvement applications.
期刊:
Neural Computing and Applications,2025年37(6):5169-5186 ISSN:0941-0643
通讯作者:
Shan Zeng
作者机构:
[Cheng Zhang; Shan Zeng; Zhiguang Yang; Hao Li] College of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan, China;[Yulong Chen] College of Medicine and Health Science, Wuhan Polytechnic University, Wuhan, China;[Yuanyan Tang] Faculty of Science and Technology, University of Macau, Macau, China
通讯机构:
[Shan Zeng] C;College of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan, China
摘要:
Intelligent transportation system (ITS) plays an important role in assisting drivers to master road information and optimize traffic flow. However, image degradation resulting from complicated environmental factors, such as motion blur caused by vehicle movement and illumination condition, caused some difficulties in current object recognition research on the ITS that may pose serious risks to driving safety. In order to tackle these challenges, this paper introduces a novel perturbation-based image super-resolution method based on GAN inversion (PSRGANI), utilizing a perturbation mechanism to better assist the latent space escape from the local optimum. In the architecture of PSRGANI, a dual encoder that preserves high-dimensional semantic feature from degraded images, extracts texture information from high-resolution (HR) images and completes the SR reconstruction process via perturbation mechanism. The additional encoder effectively decreases illuminated interference from external environment, enhancing the result robustness of SR reconstruction. A dual discriminator precisely regulates the upsampling process of the decoder for improving the quality of generated images. The additional discriminator achieves the optimal fusion of inputs from different sources by decoupling. SR experiment results reveal the higher evaluation metrics of PSRGANI in texture and decoupling compared with other SR models such as ESRGAN and DGP. In real-world ITS of traffic sign experiments, PSRGANI-applied model shows better performance (Top-1 Class Error of 3%) and faster inference speed (0.06 s per image) when compared to target detection algorithms such as YOLO and DETR. PSRGANI is demonstrated to have accurate results on degraded image recognition in terms of texture quality and evaluation metrics.
作者机构:
[Li, Mingran; Zhou, Chunhui; Tang, Wuao; Zhang, Fan] Wuhan Univ Technol, Sch Nav, Wuhan 430063, Peoples R China.;[Li, Mingran; Huang, Hongxun; Zhou, Chunhui; Tang, Wuao; Zhang, Fan] Hubei Key Lab Inland Shipping Technol, Wuhan 430063, Peoples R China.;[Huang, Liang; Wen, Yuanqiao] Wuhan Univ Technol, Intelligent Transportat Syst Res Ctr, Wuhan 430063, Peoples R China.;[Huang, Liang; Wen, Yuanqiao] Wuhan Univ Technol, Natl Engn Res Ctr Water Transport Safety, Wuhan 430063, Peoples R China.;[Huang, HX; Huang, Hongxun] Wuhan Polytech Univ, Sch Math & Comp Sci, Wuhan 430048, Peoples R China.
通讯机构:
[Huang, HX ] W;Wuhan Polytech Univ, Sch Math & Comp Sci, Wuhan 430048, Peoples R China.
关键词:
Interaction effects;Monitoring point selection;River-crossing bridge;Ship exhaust monitoring;Ship plume dispersion
摘要:
This study addresses the challenge of efficient monitoring of ship exhaust emissions in inland waterways by proposing an optimized approach to selecting bridge-based monitoring points. A micro-scale Computational Fluid Dynamics (CFD) model was developed to simulate interactions between ship exhaust plumes and river-crossing bridges, enabling precise predictions of dispersion patterns and concentrations. A numerical model incorporating pre-defined monitoring points and local environmental data was used to evaluate the influence of wind, water levels, and ship dynamics on plume behavior. The model's feasibility was validated through on-site UAV experiments. Results showed that plume dispersion is significantly affected by wind direction, wind speed, water level, and ship speed. Under extreme low water levels, the proposed three-point monitoring setup achieved a detection probability of 61.47 %, with performance improving as water levels increased. This study enhances monitoring accuracy and efficiency for riverine areas, providing a valuable tool for precise regulation of ship-emitted pollutants and supporting sustainable inland waterway management.
This study addresses the challenge of efficient monitoring of ship exhaust emissions in inland waterways by proposing an optimized approach to selecting bridge-based monitoring points. A micro-scale Computational Fluid Dynamics (CFD) model was developed to simulate interactions between ship exhaust plumes and river-crossing bridges, enabling precise predictions of dispersion patterns and concentrations. A numerical model incorporating pre-defined monitoring points and local environmental data was used to evaluate the influence of wind, water levels, and ship dynamics on plume behavior. The model's feasibility was validated through on-site UAV experiments. Results showed that plume dispersion is significantly affected by wind direction, wind speed, water level, and ship speed. Under extreme low water levels, the proposed three-point monitoring setup achieved a detection probability of 61.47 %, with performance improving as water levels increased. This study enhances monitoring accuracy and efficiency for riverine areas, providing a valuable tool for precise regulation of ship-emitted pollutants and supporting sustainable inland waterway management.
摘要:
This study investigates the fixed-time bipartite consensus (FTBC) problem for leader–follower nonlinear multi-agent systems (MASs) with nonzero control input for leader and external disturbances. The cooperative–antagonistic relationships between agents are characterized by a structurally balanced signed graph. First, a generalized Lipschitz-type condition is introduced to realize the bipartite consensus of the MASs. A distributed fixed-time controller was then designed to guarantee the FTBC. The upper bound of the convergence time and the sufficient condition for the realization of the FTBC are obtained by applying graph theory and Lyapunov theory. Finally, the effectiveness of the control algorithm is verified through numerical simulations.
作者机构:
[Zibo Huang; Fangxiu Wang; Chen Su; Yi Wang; Hui Liu] School of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan, China
摘要:
As one of the key application scenarios of wireless sensor networks, the coverage optimization of underwater wireless sensor networks (UWSNs) requires special consideration of three-dimensional spatial characteristics, which distinctly differs from traditional terrestrial environment coverage issues. To address the problems of low coverage and uneven distribution in UWSNs within a three-dimensional space, we propose a Reinforcement Learning-driven Hunter-Prey Optimization (RL-HPO) algorithm. Firstly, a nonlinear convergence factor is designed to regulate the exploration and exploitation phases, achieving an effective balance between these two stages. Secondly, by incorporating the concept of Q-learning, the algorithm can adaptively select the optimal action strategy at different stages, thereby enhancing the effectiveness of actions executed in each phase. Lastly, the Nelder-Mead simplex strategy is introduced to perturb poorly performing individuals within the population, fully exploiting their search potential and preventing the algorithm from getting trapped in local optima. The performance of the RL-HPO algorithm in three-dimensional WSN environments, both with and without obstacles, was evaluated through simulation experiments. Comparisons were made with PSO, HPO, SSA, ALGWO, and SWOA. The results demonstrate that RL-HPO significantly outperforms other algorithms in key metrics such as coverage rate, moving distance, and network connectivity. In obstacle-free scenarios, RL-HPO achieved the highest coverage rate of 96.5%, while in scenarios with obstacles, the coverage rate reached 93.3%, representing improvements of 12.58%, 13.41%, 4.15%, 8%, and 13.95% over the other algorithms, respectively.
摘要:
This brief investigates the reachable set estimation (RSE) of inertial complex-valued memristive neural networks (ICVMNNs) with bounded disturbances. By taking into account the analysis method and inequality technique, an algebraic criterion of RES is established. To deal with the inertial terms in memristive neural networks, a nonreduced-order approach is adopted. Besides, the non-separation analysis method is applied to investigate complex-valued problems. Then, a complex-valued feedback control scheme is designed to ensure that the states of ICVMNNs converge to a bounded region. Eventually, a numerical example is provided to illustrate the effectiveness of the obtained theoretical result.
关键词:
Drone-based monitoring;Construction site safety;IoT integration;GSConv;Real-time safety detection;Linformer
摘要:
Real-time safety monitoring on construction sites is essential for ensuring worker safety, but traditional detection methods face challenges in dynamic environments with moving objects, occlusions, and complex conditions. To address these limitations, we propose GS-LinYOLOv10, an improved model based on YOLOv10, specifically designed for drone-based safety monitoring. The GSConv module introduces a lightweight feature extraction mechanism, reducing computational complexity without compromising detection accuracy. The Linformer-based attention mechanism efficiently captures global context, addressing challenges in dynamic and complex environments. The model integrates IoT sensor data for real-time feedback, incorporates the GSConv module for lightweight feature extraction, and utilizes a Linformer-based attention mechanism to efficiently capture global context. These innovations reduce computational complexity while significantly improving detection accuracy. Experimental results show that GS-LinYOLOv10 achieves a precision of 91.2% and a mean average precision (mAP) of 89.4%, outperforming existing models. The integration of IoT sensors allows the drone system to dynamically adjust its monitoring focus, improving adaptability to changing environments and enhancing hazard detection. This research provides an advanced, drone-based IoT-enhanced solution for real-time construction site safety monitoring, offering a more effective and efficient approach to safety management.
Real-time safety monitoring on construction sites is essential for ensuring worker safety, but traditional detection methods face challenges in dynamic environments with moving objects, occlusions, and complex conditions. To address these limitations, we propose GS-LinYOLOv10, an improved model based on YOLOv10, specifically designed for drone-based safety monitoring. The GSConv module introduces a lightweight feature extraction mechanism, reducing computational complexity without compromising detection accuracy. The Linformer-based attention mechanism efficiently captures global context, addressing challenges in dynamic and complex environments. The model integrates IoT sensor data for real-time feedback, incorporates the GSConv module for lightweight feature extraction, and utilizes a Linformer-based attention mechanism to efficiently capture global context. These innovations reduce computational complexity while significantly improving detection accuracy. Experimental results show that GS-LinYOLOv10 achieves a precision of 91.2% and a mean average precision (mAP) of 89.4%, outperforming existing models. The integration of IoT sensors allows the drone system to dynamically adjust its monitoring focus, improving adaptability to changing environments and enhancing hazard detection. This research provides an advanced, drone-based IoT-enhanced solution for real-time construction site safety monitoring, offering a more effective and efficient approach to safety management.
期刊:
Applied Mathematics and Computation,2025年484:128994 ISSN:0096-3003
通讯作者:
Jiemei Zhao
作者机构:
[Shen, Yi; Zhao, Jiemei] School of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430023, China;[Yu, Liqi] Mathematics Department, East University of Heilongjiang, Harbin 150066, China
通讯机构:
[Jiemei Zhao] S;School of Mathematics and Computer Science, Wuhan Polytechnic University, Wuhan 430023, China
摘要:
This study is concerned with reachable set bounding of delayed second-order memristive neural networks (SMNNs) with bounded input disturbances. By applying an analytic method, some inequality techniques and an adaptive control strategy, a sufficient condition of reachable set estimation criterion is derived to guarantee that the states of delayed SMNNs are bounded by a compact ellipsoid. A non-reduced order method is employed to investigate the reachable set bounding problem instead of the reduced order method by variable substitution. In addition, the proposed result is presented in algebraic form, which is easy to test. Finally, a simulation is performed to demonstrate the validity of the proposed algorithm.
摘要:
In this article, we explain how the famous Archimedes' principle of flotation can be used to construct various floating bodies. We survey some of the most important results regarding the floating bodies, including their relations with affine surface area and projection body, their extensions in different settings such as space forms and log-concave functions, and mention some associated open problems.
摘要:
The rapid changes in the global environment have led to an unprecedented decline in biodiversity, with over 28% of species facing extinction. This includes snakes, which are key to ecological balance. Detecting snakes is challenging due to their camouflage and elusive nature, causing data loss and feature extraction difficulties in ecological monitoring. To address these challenges, we propose an enhanced snake detection model, Snake-DETR, based on RT-DETR, specifically designed for snake detection in complex natural environments. First, we designed the Enhanced Generalized Efficient Layer Aggregation Network Based on Context Anchor Attention, which enhances the feature extraction capability for occluded snakes by aggregating critical layer information and strengthening context-dependent feature extraction. Additionally, we introduced the Enhanced Feature Extraction Backbone Network Based on Context Anchor Attention, which manages input information using multiple Enhanced Generalized Efficient Layer Aggregation Networks to retain essential spatial and semantic information. Subsequently, a lightweight Group-Shuffle Convolution is used to optimize the encoder, which reduces dependency on large-scale training data, thereby making it suitable for deployment on edge devices. Finally, we incorporated the Powerful-IoU loss function to improve regression path accuracy. Experimental results on a custom dataset covering 27 snake species demonstrate that Snake-DETR achieves a good balance between model efficiency and detection performance, meeting the requirements for fine-grained snake object detection. Compared to other state-of-the-art models, Snake-DETR achieved an accuracy of 97.66%, a recall rate of 93.92%, mAP@0.5 of 95.23%, and mAP@0.5:0.95 of 72.15%, all outperforming other algorithms in the comparative tests. Furthermore, the computational load and parameter count of the model are reduced by 47.2 and 52.2%, respectively, compared to the benchmark model. Additionally, the real-time processing capability is 43.5 frames per second, meeting the demand for real-time processing. Snake-DETR demonstrates excellent performance in complex environments and is suitable for wild snake fauna monitoring and edge device deployment, providing key technical support for ecological research.