The detection of foreign object intrusion is crucial for ensuring the safety of railway operations.To address challenges such as low efficiency,suboptimal detection accuracy,and slow detection speed inherent in conven...The detection of foreign object intrusion is crucial for ensuring the safety of railway operations.To address challenges such as low efficiency,suboptimal detection accuracy,and slow detection speed inherent in conventional comprehensive video monitoring systems for railways,a railway foreign object intrusion recognition and detection system is conceived and implemented using edge computing and deep learning technologies.In a bid to raise detection accuracy,the convolutional block attention module(CBAM),including spatial and channel attention modules,is seamlessly integrated into the YOLOv5 model,giving rise to the CBAM-YOLOv5 model.Furthermore,the distance intersection-over-union_non-maximum suppression(DIo U_NMS)algorithm is employed in lieu of the weighted nonmaximum suppression algorithm,resulting in improved detection performance for intrusive targets.To accelerate detection speed,the model undergoes pruning based on the batch normalization(BN)layer,and Tensor RT inference acceleration techniques are employed,culminating in the successful deployment of the algorithm on edge devices.The CBAM-YOLOv5 model exhibits a notable 2.1%enhancement in detection accuracy when evaluated on a selfconstructed railway dataset,achieving 95.0%for mean average precision(m AP).Furthermore,the inference speed on edge devices attains a commendable 15 frame/s.展开更多
In order to solve the problem of small objects detection in unmanned aerial vehicle(UAV)aerial images with complex background,a general detection method for multi-scale small objects based on Faster region-based convo...In order to solve the problem of small objects detection in unmanned aerial vehicle(UAV)aerial images with complex background,a general detection method for multi-scale small objects based on Faster region-based convolutional neural network(Faster R-CNN)is proposed.The bird’s nest on the high-voltage tower is taken as the research object.Firstly,we use the improved convolutional neural network ResNet101 to extract object features,and then use multi-scale sliding windows to obtain the object region proposals on the convolution feature maps with different resolutions.Finally,a deconvolution operation is added to further enhance the selected feature map with higher resolution,and then it taken as a feature mapping layer of the region proposals passing to the object detection sub-network.The detection results of the bird’s nest in UAV aerial images show that the proposed method can precisely detect small objects in aerial images.展开更多
Axial compression tests were conducted on AZ31 B magnesium and A6063 aluminum thin-walled square tubes with varied lengths and induced features at different compression rates. In compression, the magnesium tubes exhib...Axial compression tests were conducted on AZ31 B magnesium and A6063 aluminum thin-walled square tubes with varied lengths and induced features at different compression rates. In compression, the magnesium tubes exhibited a "local buckling and fracture" mode, with three fracture patterns, i.e."horizontal","double-oblique", and "spiral" fractures. In general, the magnesium tube showed an inferior crashworthiness to the aluminum square tube. In addition, the effects of L/W ratio, strain rate and induced features on the crashworthiness of thin-walled square tubes were investigated. With an increase in the L/W ratio(L and W represent the tube length and width, respectively) from 1 to 4, the maximal force and global specific energy absorption decreased in a power-law trend for the magnesium tubes,while they remained approximately constant for the aluminum tubes. Furthermore, as the compression rate increased from 5×10-5 to 10 m/s, the primary crashworthiness parameters of the magnesium tubes increased in an approximately exponential manner,while for the aluminum tubes,they changed slightly. Finally,the involved induced features were proven to be not an effective method to improve the specific energy absorption of magnesium tubes, thus, more trigger types,locations,and sizes will be evaluated in future to improve the energy-absorption ability.展开更多
For the linear crack skeleton of railway bridges with irregular strike,it is difficult to accurately express the crack contour feature by using a single smoothing fitting algorithm.In order to improve the measurement ...For the linear crack skeleton of railway bridges with irregular strike,it is difficult to accurately express the crack contour feature by using a single smoothing fitting algorithm.In order to improve the measurement accuracy,a polynomial curve fitting was proposed,which used the calibration point of crack contour as the boundary point,and then put them all together to produce a continuous contour curve to achieve the crack length measurement.The method was tested by measuring the linar cracks with different shapes.It is shown that this proposed algorithm can not only solve the jagged problem generated in the crack skeleton extraction process,but also improve the crack length measurement accuracy.The relative deviation is less than 0.15,and the measurement accuracy is over 98.05%,which provides a more effective means for the crack length measurement in railway bridges.展开更多
In order to improve the steady state performance,dynamic response and power factor of traditional power factor correction(PFC)digital control method and reduce the harmonic distortion of input current,a double closed ...In order to improve the steady state performance,dynamic response and power factor of traditional power factor correction(PFC)digital control method and reduce the harmonic distortion of input current,a double closed loop active power factorcorrection(APFC)control method with feed-forward is proposed.Firstly,the small signal model of Boost PFC control systemis built and the system transfer function is deduced,and then the parameters of the main device with Boost topology is estimated.By means of the feed-forward,the system can quickly respond to the change in input voltage.Furthermore,the use ofvoltage loop and current loop can achieve input current and output voltage regulation Simulink modeling shows that this methodcan effectively control the output voltage in case of input voltage largely fluctuating,improve the system dynamic response abilityand input power factor,and reduce the input current harmonic distortion展开更多
基金supported in part by the Science and Technology Innovation Project of CHN Energy Shuo Huang Railway Development Company Ltd(No.SHTL-22-28)the Beijing Natural Science Foundation Fengtai Urban Rail Transit Frontier Research Joint Fund(No.L231002)the Major Project of China State Railway Group Co.,Ltd.(No.K2023T003)。
文摘The detection of foreign object intrusion is crucial for ensuring the safety of railway operations.To address challenges such as low efficiency,suboptimal detection accuracy,and slow detection speed inherent in conventional comprehensive video monitoring systems for railways,a railway foreign object intrusion recognition and detection system is conceived and implemented using edge computing and deep learning technologies.In a bid to raise detection accuracy,the convolutional block attention module(CBAM),including spatial and channel attention modules,is seamlessly integrated into the YOLOv5 model,giving rise to the CBAM-YOLOv5 model.Furthermore,the distance intersection-over-union_non-maximum suppression(DIo U_NMS)algorithm is employed in lieu of the weighted nonmaximum suppression algorithm,resulting in improved detection performance for intrusive targets.To accelerate detection speed,the model undergoes pruning based on the batch normalization(BN)layer,and Tensor RT inference acceleration techniques are employed,culminating in the successful deployment of the algorithm on edge devices.The CBAM-YOLOv5 model exhibits a notable 2.1%enhancement in detection accuracy when evaluated on a selfconstructed railway dataset,achieving 95.0%for mean average precision(m AP).Furthermore,the inference speed on edge devices attains a commendable 15 frame/s.
基金National Defense Pre-research Fund Project(No.KMGY318002531)。
文摘In order to solve the problem of small objects detection in unmanned aerial vehicle(UAV)aerial images with complex background,a general detection method for multi-scale small objects based on Faster region-based convolutional neural network(Faster R-CNN)is proposed.The bird’s nest on the high-voltage tower is taken as the research object.Firstly,we use the improved convolutional neural network ResNet101 to extract object features,and then use multi-scale sliding windows to obtain the object region proposals on the convolution feature maps with different resolutions.Finally,a deconvolution operation is added to further enhance the selected feature map with higher resolution,and then it taken as a feature mapping layer of the region proposals passing to the object detection sub-network.The detection results of the bird’s nest in UAV aerial images show that the proposed method can precisely detect small objects in aerial images.
基金Project(2017JBM041)supported by the Fundamental Research Funds for the Central Universities,ChinaProject(51505024)supported by the National Natural Science Foundation of ChinaProject supported by the Foundation of Zhejiang Key Laboratory of Automobile Safety Technology of China
文摘Axial compression tests were conducted on AZ31 B magnesium and A6063 aluminum thin-walled square tubes with varied lengths and induced features at different compression rates. In compression, the magnesium tubes exhibited a "local buckling and fracture" mode, with three fracture patterns, i.e."horizontal","double-oblique", and "spiral" fractures. In general, the magnesium tube showed an inferior crashworthiness to the aluminum square tube. In addition, the effects of L/W ratio, strain rate and induced features on the crashworthiness of thin-walled square tubes were investigated. With an increase in the L/W ratio(L and W represent the tube length and width, respectively) from 1 to 4, the maximal force and global specific energy absorption decreased in a power-law trend for the magnesium tubes,while they remained approximately constant for the aluminum tubes. Furthermore, as the compression rate increased from 5×10-5 to 10 m/s, the primary crashworthiness parameters of the magnesium tubes increased in an approximately exponential manner,while for the aluminum tubes,they changed slightly. Finally,the involved induced features were proven to be not an effective method to improve the specific energy absorption of magnesium tubes, thus, more trigger types,locations,and sizes will be evaluated in future to improve the energy-absorption ability.
基金National Defense Pre-Research Fund Project(No.060601)Wanqiao Education Fund Project(No.06010023)。
文摘For the linear crack skeleton of railway bridges with irregular strike,it is difficult to accurately express the crack contour feature by using a single smoothing fitting algorithm.In order to improve the measurement accuracy,a polynomial curve fitting was proposed,which used the calibration point of crack contour as the boundary point,and then put them all together to produce a continuous contour curve to achieve the crack length measurement.The method was tested by measuring the linar cracks with different shapes.It is shown that this proposed algorithm can not only solve the jagged problem generated in the crack skeleton extraction process,but also improve the crack length measurement accuracy.The relative deviation is less than 0.15,and the measurement accuracy is over 98.05%,which provides a more effective means for the crack length measurement in railway bridges.
基金National Natural Science Foundation of China(No.61261029)
文摘In order to improve the steady state performance,dynamic response and power factor of traditional power factor correction(PFC)digital control method and reduce the harmonic distortion of input current,a double closed loop active power factorcorrection(APFC)control method with feed-forward is proposed.Firstly,the small signal model of Boost PFC control systemis built and the system transfer function is deduced,and then the parameters of the main device with Boost topology is estimated.By means of the feed-forward,the system can quickly respond to the change in input voltage.Furthermore,the use ofvoltage loop and current loop can achieve input current and output voltage regulation Simulink modeling shows that this methodcan effectively control the output voltage in case of input voltage largely fluctuating,improve the system dynamic response abilityand input power factor,and reduce the input current harmonic distortion
