This paper introduces an optimized planning approach for integrating photovoltaic as distributed generation (PV-DG) into the radial distribution power systems, utilizing exhaustive load flow (ELF), loss sensitivity fa...This paper introduces an optimized planning approach for integrating photovoltaic as distributed generation (PV-DG) into the radial distribution power systems, utilizing exhaustive load flow (ELF), loss sensitivity factor (LSF), genetic algorithms (GA) methods, and numerical method based on LSF. The methodology aims to determine the optimal allocation and sizing of multiple PV-DG to minimize power loss through time series power flow analysis. An approach utilizing continuous sensitivity analysis is developed and inherently leverages power flow and loss equations to compute LSF of all buses in the system towards employing a dynamic PV-DG model for more accurate results. The algorithm uses a numerical grid search method to optimize PV-DG placement in a power distribution system, focusing on minimizing system losses. It combines iterative analysis, sensitivity assessment, and comprehensive visualization to identify and present the optimal PV-DG configurations. The present-ed algorithms are verified through co-simulation framework combining MATLAB and OpenDSS to carry out analysis for 12-bus radial distribution test system. The proposed numerical method is compared with other algorithms, such as ELF, LSF methods, and Genetic Algorithms (GA). Results show that the proposed numerical method performs well in comparison with LSF and ELF solutions.展开更多
With the continuous growth of power demand and the diversification of power consumption structure,the loss of distribution network has gradually become the focus of attention.Given the problems of single loss reductio...With the continuous growth of power demand and the diversification of power consumption structure,the loss of distribution network has gradually become the focus of attention.Given the problems of single loss reduction measure,lack of economy,and practicality in existing research,this paper proposes an optimization method of distribution network loss reduction based on tabu search algorithm and optimizes the combination and parameter configuration of loss reduction measure.The optimization model is developed with the goal of maximizing comprehensive benefits,incorporating both economic and environmental factors,and accounting for investment costs,including the loss of power reduction.Additionally,the model ensures that constraint conditions such as power flow equations,voltage deviations,and line transmission capacities are satisfied.The solution is obtained through a tabu search algorithm,which is well-suited for solving nonlinear problems with multiple constraints.Combined with the example of 10kV25 node construction,the simulation results show that the method can significantly reduce the network loss on the basis of ensuring the economy and environmental protection of the system,which provides a theoretical basis for distribution network planning.展开更多
Non-technical losses(NTL)of electric power are a serious problem for electric distribution companies.The solution determines the cost,stability,reliability,and quality of the supplied electricity.The widespread use of...Non-technical losses(NTL)of electric power are a serious problem for electric distribution companies.The solution determines the cost,stability,reliability,and quality of the supplied electricity.The widespread use of advanced metering infrastructure(AMI)and Smart Grid allows all participants in the distribution grid to store and track electricity consumption.During the research,a machine learning model is developed that allows analyzing and predicting the probability of NTL for each consumer of the distribution grid based on daily electricity consumption readings.This model is an ensemble meta-algorithm(stacking)that generalizes the algorithms of random forest,LightGBM,and a homogeneous ensemble of artificial neural networks.The best accuracy of the proposed meta-algorithm in comparison to basic classifiers is experimentally confirmed on the test sample.Such a model,due to good accuracy indicators(ROC-AUC-0.88),can be used as a methodological basis for a decision support system,the purpose of which is to form a sample of suspected NTL sources.The use of such a sample will allow the top management of electric distribution companies to increase the efficiency of raids by performers,making them targeted and accurate,which should contribute to the fight against NTL and the sustainable development of the electric power industry.展开更多
Designing appropriate loss functions is critical to the success of supervised learning models.However,most conventional losses are fixed and manually designed,making them suboptimal for diverse and dynamic learning sc...Designing appropriate loss functions is critical to the success of supervised learning models.However,most conventional losses are fixed and manually designed,making them suboptimal for diverse and dynamic learning scenarios.In this work,we propose an Adaptive Meta-Loss Network(Adaptive-MLN)that learns to generate taskagnostic loss functions tailored to evolving classification problems.Unlike traditional methods that rely on static objectives,Adaptive-MLN treats the loss function itself as a trainable component,parameterized by a shallow neural network.To enable flexible,gradient-free optimization,we introduce a hybrid evolutionary approach that combines GeneticAlgorithms(GA)for global exploration and Evolution Strategies(ES)for local refinement.This co-evolutionary process dynamically adjusts the loss landscape,improvingmodel generalization without relying on analytic gradients or handcrafted heuristics.Experimental evaluations on synthetic tasks and the CIFAR-10 andMNIST datasets demonstrate that our approach consistently outperforms standard losses such as Cross-Entropy and Mean Squared Error in terms of accuracy,convergence,and adaptability.展开更多
Concrete-filled steel tubes(CFST)are widely utilized in civil engineering due to their superior load-bearing capacity,ductility,and seismic resistance.However,existing design codes,such as AISC and Eurocode 4,tend to ...Concrete-filled steel tubes(CFST)are widely utilized in civil engineering due to their superior load-bearing capacity,ductility,and seismic resistance.However,existing design codes,such as AISC and Eurocode 4,tend to be excessively conservative as they fail to account for the composite action between the steel tube and the concrete core.To address this limitation,this study proposes a hybrid model that integrates XGBoost with the Pied Kingfisher Optimizer(PKO),a nature-inspired algorithm,to enhance the accuracy of shear strength prediction for CFST columns.Additionally,quantile regression is employed to construct prediction intervals for the ultimate shear force,while the Asymmetric Squared Error Loss(ASEL)function is incorporated to mitigate overestimation errors.The computational results demonstrate that the PKO-XGBoost model delivers superior predictive accuracy,achieving a Mean Absolute Percentage Error(MAPE)of 4.431%and R2 of 0.9925 on the test set.Furthermore,the ASEL-PKO-XGBoost model substantially reduces overestimation errors to 28.26%,with negligible impact on predictive performance.Additionally,based on the Genetic Algorithm(GA)and existing equation models,a strength equation model is developed,achieving markedly higher accuracy than existing models(R^(2)=0.934).Lastly,web-based Graphical User Interfaces(GUIs)were developed to enable real-time prediction.展开更多
With the increasing penetration of renewable energy,the coordination of energy storage with thermal power for frequency regulation has become an effective means to enhance grid frequency security.Addressing the challe...With the increasing penetration of renewable energy,the coordination of energy storage with thermal power for frequency regulation has become an effective means to enhance grid frequency security.Addressing the challenge of improving the frequency regulation performance of a thermal-storage primary frequency regulation system while reducing its associated losses,this paper proposes a multi-dimensional cooperative optimization strategy for the control parameters of a combined thermal-storage system,considering regulation losses.First,the frequency regulation losses of various components within the thermal power unit are quantified,and a calculation method for energy storage regulation loss is proposed,based on Depth of Discharge(DOD)and C-rate.Second,a thermal-storage cooperative control method based on series compensation is developed to improve the system’s frequency regulation performance.Third,targeting system regulation loss cost and regulation output,and considering constraints on output overshoot and system parameters,an improved Particle Swarm Optimization(PSO)algorithm is employed to tune the parameters of the low-pass filter and the series compensator,thereby reducing regulation losses while enhancing performance.Finally,simulation results demonstrate that the total loss cost of the proposed control strategy is comparable to that of a system with only thermal power participation.However,the thermal power loss cost is reduced by 42.16%compared to the thermal-only case,while simultaneously improving system frequency stability.Thus,the proposed strategy effectively balances system frequency stability and economic efficiency.展开更多
Workers’exposure to excessive noise is a big universal work-related challenges.One of the major consequences of exposure to noise is permanent or transient hearing loss.The current study sought to utilize audiometric...Workers’exposure to excessive noise is a big universal work-related challenges.One of the major consequences of exposure to noise is permanent or transient hearing loss.The current study sought to utilize audiometric data to weigh and prioritize the factors affecting workers’hearing loss based using the Support Vector Machine(SVM)algorithm.This cross sectional-descriptive study was conducted in 2017 in a mining industry in southeast Iran.The participating workers(n=150)were divided into three groups of 50 based on the sound pressure level to which they were exposed(two experimental groups and one control group).Audiometric tests were carried out for all members of each group.The study generally entailed the following steps:(1)selecting predicting variables to weigh and prioritize factors affecting hearing loss;(2)conducting audiometric tests and assessing permanent hearing loss in each ear and then evaluating total hearing loss;(3)categorizing different types of hearing loss;(4)weighing and prioritizing factors that affect hearing loss based on the SVM algorithm;and(5)assessing the error rate and accuracy of the models.The collected data were fed into SPSS 18,followed by conducting linear regression and paired samples t-test.It was revealed that,in the first model(SPL<70 dBA),the frequency of 8 KHz had the greatest impact(with a weight of 33%),while noise had the smallest influence(with a weight of 5%).The accuracy of this model was 100%.In the second model(70<SPL<80 dBA),the frequency of 4 KHz had the most profound effect(with a weight of 21%),whereas the frequency of 250 Hz had the lowest impact(with a weight of 6%).The accuracy of this model was 100%too.In the third model(SPL>85 dBA),the frequency of 4 KHz had the highest impact(with a weight of 22%),while the frequency of 250 Hz had the smallest influence(with a weight of 3%).The accuracy of this model was 100%too.In the fourth model,the frequency of 4 KHz had the greatest effect(with a weight of 24%),while the frequency of 500 Hz had the smallest effect(with a weight of 4%).The accuracy of this model was found to be 94%.According to the modeling conducted using the SVM algorithm,the frequency of 4 KHz has the most profound effect on predicting changes in hearing loss.Given the high accuracy of the obtained model,this algorithm is an appropriate and powerful tool to predict and model hearing loss.展开更多
This work presents a fuzzy based methodology for distribution system feeder reconfiguration considering DSTATCOM with an objective of minimizing real power loss and operating cost. Installation costs of DSTATCOM devic...This work presents a fuzzy based methodology for distribution system feeder reconfiguration considering DSTATCOM with an objective of minimizing real power loss and operating cost. Installation costs of DSTATCOM devices and the cost of system operation, namely, energy loss cost due to both reconfiguration and DSTATCOM placement, are combined to form the objective function to be minimized. The distribution system tie switches, DSTATCOM location and size have been optimally determined to obtain an appropriate operational condition. In the proposed approach, the fuzzy membership function of loss sensitivity is used for the selection of weak nodes in the power system for the placement of DSTATCOM and the optimal parameter settings of the DFACTS device along with optimal selection of tie switches in reconfiguration process are governed by genetic algorithm(GA). Simulation results on IEEE 33-bus and IEEE 69-bus test systems concluded that the combinatorial method using DSTATCOM and reconfiguration is preferable to reduce power losses to 34.44% for 33-bus system and to 45.43% for 69-bus system.展开更多
This paper presents a powerful approach to find the optimal size and location of distributed generation units in a distribution system using GA(Genetic Optimization algorithm).It is proved that GA method is fast and e...This paper presents a powerful approach to find the optimal size and location of distributed generation units in a distribution system using GA(Genetic Optimization algorithm).It is proved that GA method is fast and easy tool to enable the planners to select accurate and the optimum size of generators to improve the system voltage profile in addition to reduce the active and reactive power loss.GA fitness function is introduced including the active power losses,reactive power losses and the cumulative voltage deviation variables with selecting weight of each variable.GA fitness function is subjected to voltage constraints,active and reactive power losses constraints and DG size constraint.展开更多
In smart driving for rail transit,a reliable obstacle detection system is an important guarantee for the safety of trains.Therein,the detection of the rail area directly affects the accuracy of the system to identify ...In smart driving for rail transit,a reliable obstacle detection system is an important guarantee for the safety of trains.Therein,the detection of the rail area directly affects the accuracy of the system to identify dangerous targets.Both the rail line and the lane are presented as thin line shapes in the image,but the rail scene is more complex,and the color of the rail line is more difficult to distinguish from the background.By comparison,there are already many deep learning-based lane detection algorithms,but there is a lack of public datasets and targeted deep learning detection algorithms for rail line detection.To address this,this paper constructs a rail image dataset RailwayLine and labels the rail line for the training and testing of models.This dataset contains rich rail images including single-rail,multi-rail,straight rail,curved rail,crossing rails,occlusion,blur,and different lighting conditions.To address the problem of the lack of deep learning-based rail line detection algorithms,we improve the CLRNet algorithm which has an excellent performance in lane detection,and propose the CLRNet-R algorithm for rail line detection.To address the problem of the rail line being thin and occupying fewer pixels in the image,making it difficult to distinguish from complex backgrounds,we introduce an attention mechanism to enhance global feature extraction ability and add a semantic segmentation head to enhance the features of the rail region by the binary probability of rail lines.To address the poor curve recognition performance and unsmooth output lines in the original CLRNet algorithm,we improve the weight allocation for line intersection-over-union calculation in the original framework and propose two loss functions based on local slopes to optimize the model’s local sampling point training constraints,improving the model’s fitting performance on curved rails and obtaining smooth and stable rail line detection results.Through experiments,this paper demonstrates that compared with other mainstream lane detection algorithms,the algorithm proposed in this paper has a better performance for rail line detection.展开更多
针对车辆漆面缺陷检测精度低、检测算法参数量大、难易样本不均匀等问题,提出一种基于改进YOLOv8的车辆漆面检测算法。首先,为了提升划痕状缺陷检测能力并降低模型规模,将DAT(Deformable Attention Transformer)注意力机制引入主干网络...针对车辆漆面缺陷检测精度低、检测算法参数量大、难易样本不均匀等问题,提出一种基于改进YOLOv8的车辆漆面检测算法。首先,为了提升划痕状缺陷检测能力并降低模型规模,将DAT(Deformable Attention Transformer)注意力机制引入主干网络来增强长距离特征依赖关系,同时使用幻影卷积(GhostConv)替换网络中的卷积(Conv)模块。然后,为了提升特征提取能力并进一步降低模型规模,结合FasterBlock模块与高效多尺度注意力(EMA)机制提出C2f-E(C2f Based on EMA)模块。接着,为了提高小目标检测性能,基于双向特征金字塔网络(BiFPN)进行设计,并增加小目标检测头与多尺度特征融合支路,提出BiFPN-D(BiFPN with Small Object Detection Head)颈部金字塔结构。最后,为了解决难易样本的平衡问题并提高针对小目标缺陷的检测性能,使用WIoUv3(Wise-Intersection over Union version 3)作为训练网络的损失函数。在自建的车辆漆面缺陷数据集上进行训练并开展对比实验。实验结果表明,相较于YOLOv8n,改进模型的均值平均精度(mAP@0.5)提高了5.5百分点、规模减小了1.4×106。展开更多
文摘This paper introduces an optimized planning approach for integrating photovoltaic as distributed generation (PV-DG) into the radial distribution power systems, utilizing exhaustive load flow (ELF), loss sensitivity factor (LSF), genetic algorithms (GA) methods, and numerical method based on LSF. The methodology aims to determine the optimal allocation and sizing of multiple PV-DG to minimize power loss through time series power flow analysis. An approach utilizing continuous sensitivity analysis is developed and inherently leverages power flow and loss equations to compute LSF of all buses in the system towards employing a dynamic PV-DG model for more accurate results. The algorithm uses a numerical grid search method to optimize PV-DG placement in a power distribution system, focusing on minimizing system losses. It combines iterative analysis, sensitivity assessment, and comprehensive visualization to identify and present the optimal PV-DG configurations. The present-ed algorithms are verified through co-simulation framework combining MATLAB and OpenDSS to carry out analysis for 12-bus radial distribution test system. The proposed numerical method is compared with other algorithms, such as ELF, LSF methods, and Genetic Algorithms (GA). Results show that the proposed numerical method performs well in comparison with LSF and ELF solutions.
文摘With the continuous growth of power demand and the diversification of power consumption structure,the loss of distribution network has gradually become the focus of attention.Given the problems of single loss reduction measure,lack of economy,and practicality in existing research,this paper proposes an optimization method of distribution network loss reduction based on tabu search algorithm and optimizes the combination and parameter configuration of loss reduction measure.The optimization model is developed with the goal of maximizing comprehensive benefits,incorporating both economic and environmental factors,and accounting for investment costs,including the loss of power reduction.Additionally,the model ensures that constraint conditions such as power flow equations,voltage deviations,and line transmission capacities are satisfied.The solution is obtained through a tabu search algorithm,which is well-suited for solving nonlinear problems with multiple constraints.Combined with the example of 10kV25 node construction,the simulation results show that the method can significantly reduce the network loss on the basis of ensuring the economy and environmental protection of the system,which provides a theoretical basis for distribution network planning.
文摘Non-technical losses(NTL)of electric power are a serious problem for electric distribution companies.The solution determines the cost,stability,reliability,and quality of the supplied electricity.The widespread use of advanced metering infrastructure(AMI)and Smart Grid allows all participants in the distribution grid to store and track electricity consumption.During the research,a machine learning model is developed that allows analyzing and predicting the probability of NTL for each consumer of the distribution grid based on daily electricity consumption readings.This model is an ensemble meta-algorithm(stacking)that generalizes the algorithms of random forest,LightGBM,and a homogeneous ensemble of artificial neural networks.The best accuracy of the proposed meta-algorithm in comparison to basic classifiers is experimentally confirmed on the test sample.Such a model,due to good accuracy indicators(ROC-AUC-0.88),can be used as a methodological basis for a decision support system,the purpose of which is to form a sample of suspected NTL sources.The use of such a sample will allow the top management of electric distribution companies to increase the efficiency of raids by performers,making them targeted and accurate,which should contribute to the fight against NTL and the sustainable development of the electric power industry.
基金supported by the National Natural Science Foundation of China(NSFC)under Grant number:82171965.
文摘Designing appropriate loss functions is critical to the success of supervised learning models.However,most conventional losses are fixed and manually designed,making them suboptimal for diverse and dynamic learning scenarios.In this work,we propose an Adaptive Meta-Loss Network(Adaptive-MLN)that learns to generate taskagnostic loss functions tailored to evolving classification problems.Unlike traditional methods that rely on static objectives,Adaptive-MLN treats the loss function itself as a trainable component,parameterized by a shallow neural network.To enable flexible,gradient-free optimization,we introduce a hybrid evolutionary approach that combines GeneticAlgorithms(GA)for global exploration and Evolution Strategies(ES)for local refinement.This co-evolutionary process dynamically adjusts the loss landscape,improvingmodel generalization without relying on analytic gradients or handcrafted heuristics.Experimental evaluations on synthetic tasks and the CIFAR-10 andMNIST datasets demonstrate that our approach consistently outperforms standard losses such as Cross-Entropy and Mean Squared Error in terms of accuracy,convergence,and adaptability.
基金funded by United Arab Emirates University(UAEU)under the UAEU-AUA grant number G00004577(12N145)with the corresponding grant at Universiti Malaya(UM)under grant number IF019-2024.
文摘Concrete-filled steel tubes(CFST)are widely utilized in civil engineering due to their superior load-bearing capacity,ductility,and seismic resistance.However,existing design codes,such as AISC and Eurocode 4,tend to be excessively conservative as they fail to account for the composite action between the steel tube and the concrete core.To address this limitation,this study proposes a hybrid model that integrates XGBoost with the Pied Kingfisher Optimizer(PKO),a nature-inspired algorithm,to enhance the accuracy of shear strength prediction for CFST columns.Additionally,quantile regression is employed to construct prediction intervals for the ultimate shear force,while the Asymmetric Squared Error Loss(ASEL)function is incorporated to mitigate overestimation errors.The computational results demonstrate that the PKO-XGBoost model delivers superior predictive accuracy,achieving a Mean Absolute Percentage Error(MAPE)of 4.431%and R2 of 0.9925 on the test set.Furthermore,the ASEL-PKO-XGBoost model substantially reduces overestimation errors to 28.26%,with negligible impact on predictive performance.Additionally,based on the Genetic Algorithm(GA)and existing equation models,a strength equation model is developed,achieving markedly higher accuracy than existing models(R^(2)=0.934).Lastly,web-based Graphical User Interfaces(GUIs)were developed to enable real-time prediction.
基金supported by the Science and Technology Development Project of Jilin Province(Project No.YDZJ202301ZYTS284).
文摘With the increasing penetration of renewable energy,the coordination of energy storage with thermal power for frequency regulation has become an effective means to enhance grid frequency security.Addressing the challenge of improving the frequency regulation performance of a thermal-storage primary frequency regulation system while reducing its associated losses,this paper proposes a multi-dimensional cooperative optimization strategy for the control parameters of a combined thermal-storage system,considering regulation losses.First,the frequency regulation losses of various components within the thermal power unit are quantified,and a calculation method for energy storage regulation loss is proposed,based on Depth of Discharge(DOD)and C-rate.Second,a thermal-storage cooperative control method based on series compensation is developed to improve the system’s frequency regulation performance.Third,targeting system regulation loss cost and regulation output,and considering constraints on output overshoot and system parameters,an improved Particle Swarm Optimization(PSO)algorithm is employed to tune the parameters of the low-pass filter and the series compensator,thereby reducing regulation losses while enhancing performance.Finally,simulation results demonstrate that the total loss cost of the proposed control strategy is comparable to that of a system with only thermal power participation.However,the thermal power loss cost is reduced by 42.16%compared to the thermal-only case,while simultaneously improving system frequency stability.Thus,the proposed strategy effectively balances system frequency stability and economic efficiency.
基金This study stemmed from a research project(code number:96000838)which was sponsored by the Institute for Futures Studies in Health at Kerman University of Medical Sciences.
文摘Workers’exposure to excessive noise is a big universal work-related challenges.One of the major consequences of exposure to noise is permanent or transient hearing loss.The current study sought to utilize audiometric data to weigh and prioritize the factors affecting workers’hearing loss based using the Support Vector Machine(SVM)algorithm.This cross sectional-descriptive study was conducted in 2017 in a mining industry in southeast Iran.The participating workers(n=150)were divided into three groups of 50 based on the sound pressure level to which they were exposed(two experimental groups and one control group).Audiometric tests were carried out for all members of each group.The study generally entailed the following steps:(1)selecting predicting variables to weigh and prioritize factors affecting hearing loss;(2)conducting audiometric tests and assessing permanent hearing loss in each ear and then evaluating total hearing loss;(3)categorizing different types of hearing loss;(4)weighing and prioritizing factors that affect hearing loss based on the SVM algorithm;and(5)assessing the error rate and accuracy of the models.The collected data were fed into SPSS 18,followed by conducting linear regression and paired samples t-test.It was revealed that,in the first model(SPL<70 dBA),the frequency of 8 KHz had the greatest impact(with a weight of 33%),while noise had the smallest influence(with a weight of 5%).The accuracy of this model was 100%.In the second model(70<SPL<80 dBA),the frequency of 4 KHz had the most profound effect(with a weight of 21%),whereas the frequency of 250 Hz had the lowest impact(with a weight of 6%).The accuracy of this model was 100%too.In the third model(SPL>85 dBA),the frequency of 4 KHz had the highest impact(with a weight of 22%),while the frequency of 250 Hz had the smallest influence(with a weight of 3%).The accuracy of this model was 100%too.In the fourth model,the frequency of 4 KHz had the greatest effect(with a weight of 24%),while the frequency of 500 Hz had the smallest effect(with a weight of 4%).The accuracy of this model was found to be 94%.According to the modeling conducted using the SVM algorithm,the frequency of 4 KHz has the most profound effect on predicting changes in hearing loss.Given the high accuracy of the obtained model,this algorithm is an appropriate and powerful tool to predict and model hearing loss.
基金supported by Borujerd Branch,Islamic Azad University Iran
文摘This work presents a fuzzy based methodology for distribution system feeder reconfiguration considering DSTATCOM with an objective of minimizing real power loss and operating cost. Installation costs of DSTATCOM devices and the cost of system operation, namely, energy loss cost due to both reconfiguration and DSTATCOM placement, are combined to form the objective function to be minimized. The distribution system tie switches, DSTATCOM location and size have been optimally determined to obtain an appropriate operational condition. In the proposed approach, the fuzzy membership function of loss sensitivity is used for the selection of weak nodes in the power system for the placement of DSTATCOM and the optimal parameter settings of the DFACTS device along with optimal selection of tie switches in reconfiguration process are governed by genetic algorithm(GA). Simulation results on IEEE 33-bus and IEEE 69-bus test systems concluded that the combinatorial method using DSTATCOM and reconfiguration is preferable to reduce power losses to 34.44% for 33-bus system and to 45.43% for 69-bus system.
文摘This paper presents a powerful approach to find the optimal size and location of distributed generation units in a distribution system using GA(Genetic Optimization algorithm).It is proved that GA method is fast and easy tool to enable the planners to select accurate and the optimum size of generators to improve the system voltage profile in addition to reduce the active and reactive power loss.GA fitness function is introduced including the active power losses,reactive power losses and the cumulative voltage deviation variables with selecting weight of each variable.GA fitness function is subjected to voltage constraints,active and reactive power losses constraints and DG size constraint.
基金the Sichuan Science and Technology Program(No.2022YFS0557)the National Natural Science Foundation of China(No.61972271)。
文摘In smart driving for rail transit,a reliable obstacle detection system is an important guarantee for the safety of trains.Therein,the detection of the rail area directly affects the accuracy of the system to identify dangerous targets.Both the rail line and the lane are presented as thin line shapes in the image,but the rail scene is more complex,and the color of the rail line is more difficult to distinguish from the background.By comparison,there are already many deep learning-based lane detection algorithms,but there is a lack of public datasets and targeted deep learning detection algorithms for rail line detection.To address this,this paper constructs a rail image dataset RailwayLine and labels the rail line for the training and testing of models.This dataset contains rich rail images including single-rail,multi-rail,straight rail,curved rail,crossing rails,occlusion,blur,and different lighting conditions.To address the problem of the lack of deep learning-based rail line detection algorithms,we improve the CLRNet algorithm which has an excellent performance in lane detection,and propose the CLRNet-R algorithm for rail line detection.To address the problem of the rail line being thin and occupying fewer pixels in the image,making it difficult to distinguish from complex backgrounds,we introduce an attention mechanism to enhance global feature extraction ability and add a semantic segmentation head to enhance the features of the rail region by the binary probability of rail lines.To address the poor curve recognition performance and unsmooth output lines in the original CLRNet algorithm,we improve the weight allocation for line intersection-over-union calculation in the original framework and propose two loss functions based on local slopes to optimize the model’s local sampling point training constraints,improving the model’s fitting performance on curved rails and obtaining smooth and stable rail line detection results.Through experiments,this paper demonstrates that compared with other mainstream lane detection algorithms,the algorithm proposed in this paper has a better performance for rail line detection.
文摘针对车辆漆面缺陷检测精度低、检测算法参数量大、难易样本不均匀等问题,提出一种基于改进YOLOv8的车辆漆面检测算法。首先,为了提升划痕状缺陷检测能力并降低模型规模,将DAT(Deformable Attention Transformer)注意力机制引入主干网络来增强长距离特征依赖关系,同时使用幻影卷积(GhostConv)替换网络中的卷积(Conv)模块。然后,为了提升特征提取能力并进一步降低模型规模,结合FasterBlock模块与高效多尺度注意力(EMA)机制提出C2f-E(C2f Based on EMA)模块。接着,为了提高小目标检测性能,基于双向特征金字塔网络(BiFPN)进行设计,并增加小目标检测头与多尺度特征融合支路,提出BiFPN-D(BiFPN with Small Object Detection Head)颈部金字塔结构。最后,为了解决难易样本的平衡问题并提高针对小目标缺陷的检测性能,使用WIoUv3(Wise-Intersection over Union version 3)作为训练网络的损失函数。在自建的车辆漆面缺陷数据集上进行训练并开展对比实验。实验结果表明,相较于YOLOv8n,改进模型的均值平均精度(mAP@0.5)提高了5.5百分点、规模减小了1.4×106。
