Forging spur gears are widely used in the driving system of mining machinery and equipment due to their higher strength and dimensional accuracy.For the purpose of precisely calculating the volume of cylindrical spur ...Forging spur gears are widely used in the driving system of mining machinery and equipment due to their higher strength and dimensional accuracy.For the purpose of precisely calculating the volume of cylindrical spur gear billet in cold precision forging,a new theoretical method named average circle method was put forward.With this method,a series of gear billet volumes were calculated.Comparing with the accurate three-dimensional modeling method,the accuracy of average circle method by theoretical calculation was estimated and the maximum relative error of average circle method was less than 1.5%,which was in good agreement with the experimental results.Relative errors of the calculated and the experimental for obtaining the gear billet volumes with reference circle method are larger than those of the average circle method.It shows that average circle method possesses a higher calculation accuracy than reference circle method(traditional method),which should be worth popularizing widely in calculation of spur gear billet volume.展开更多
Landmines continue to pose an ongoing threat in various regions around the world,with countless buried landmines affecting numerous human lives.The detonation of these landmines results in thousands of casualties repo...Landmines continue to pose an ongoing threat in various regions around the world,with countless buried landmines affecting numerous human lives.The detonation of these landmines results in thousands of casualties reported worldwide annually.Therefore,there is a pressing need to employ diverse landmine detection techniques for their removal.One effective approach for landmine detection is UAV(Unmanned Aerial Vehicle)based AirborneMagnetometry,which identifies magnetic anomalies in the local terrestrial magnetic field.It can generate a contour plot or heat map that visually represents the magnetic field strength.Despite the effectiveness of this approach,landmine removal remains a challenging and resource-intensive task,fraughtwith risks.Edge computing,on the other hand,can play a crucial role in critical drone monitoring applications like landmine detection.By processing data locally on a nearby edge server,edge computing can reduce communication latency and bandwidth requirements,allowing real-time analysis of magnetic field data.It enables faster decision-making and more efficient landmine detection,potentially saving lives and minimizing the risks involved in the process.Furthermore,edge computing can provide enhanced security and privacy by keeping sensitive data close to the source,reducing the chances of data exposure during transmission.This paper introduces the MAGnetometry Imaging based Classification System(MAGICS),a fully automated UAV-based system designed for landmine and buried object detection and localization.We have developed an efficient deep learning-based strategy for automatic image classification using magnetometry dataset traces.By simulating the proposal in various network scenarios,we have successfully detected landmine signatures present in themagnetometry images.The trained models exhibit significant performance improvements,achieving a maximum mean average precision value of 97.8%.展开更多
In this paper, we discuss the precise asymptotics of moving-average process Xt =∞∑j=0 ajEt-j under some suitable conditions, where {εt, t∈ Z} is a sequence j=0 of stationary ALNQD random variables with mean zeros...In this paper, we discuss the precise asymptotics of moving-average process Xt =∞∑j=0 ajEt-j under some suitable conditions, where {εt, t∈ Z} is a sequence j=0 of stationary ALNQD random variables with mean zeros and finite variances.展开更多
为解决传统目标检测算法在地铁、商场以及交通堵塞等地区高密度人群中因目标重叠和尺寸偏小而难以检测的问题,文中提出一种基于YOLOv5(You Only Look Once version 5)网络的目标检测算法。在算法模型的锚框部分引入新特征图来设计添加...为解决传统目标检测算法在地铁、商场以及交通堵塞等地区高密度人群中因目标重叠和尺寸偏小而难以检测的问题,文中提出一种基于YOLOv5(You Only Look Once version 5)网络的目标检测算法。在算法模型的锚框部分引入新特征图来设计添加小目标检测层,以此提升检测小目标的准确性。通过重新定义一个卷积层,在YOLOv5中添加SOCA(Second-Order Channel Attention)注意力机制,提高了模型对复杂场景和遮挡的鲁棒性。引入Focal_EIoU(Focal and Efficient Intersection over Union)替换原始模型的损失函数CIoU(Complete Intersection over Union),从而提高了模型在高密度目标上的检测精度。实验结果表明,改进YOLOv5算法在CrowdHuman数据集上的平均检测精度比原模型提高了6.7百分点,召回率提高了3.8百分点,优于FPN(Feature Pyramid Network)和RetinaNet算法,实现了对高密度人群的目标检测优化。展开更多
为解决在目标检测网络中使用特征融合方法带来的参数量大、计算复杂度高的问题,提出了一种融合无参注意力机制(SimAM)的特征融合方法。对动态蛇形卷积(DSConv)进行轻量化处理(Light-DSConv)。利用该结构自主学习目标几何形状的能力,对...为解决在目标检测网络中使用特征融合方法带来的参数量大、计算复杂度高的问题,提出了一种融合无参注意力机制(SimAM)的特征融合方法。对动态蛇形卷积(DSConv)进行轻量化处理(Light-DSConv)。利用该结构自主学习目标几何形状的能力,对小目标的特征进行二次提取。利用SimAM模块对特征图空间域的重要性进行划分并与通道域权重相结合,进一步提升模型性能。在Pascal VOC 2007测试集上测试融合模块的有效性。结果表明:轻量化后,单个DSConv结构参数量下降85.6%。模型平均精度(mean average precision,mAP)比基线模型增加了4.41%,比添加现有特征融合方法模型平均增加3.78%。所提出模块的参数量、计算量、检测速度与现阶段其它方法相比均具有一定优势。展开更多
文摘Forging spur gears are widely used in the driving system of mining machinery and equipment due to their higher strength and dimensional accuracy.For the purpose of precisely calculating the volume of cylindrical spur gear billet in cold precision forging,a new theoretical method named average circle method was put forward.With this method,a series of gear billet volumes were calculated.Comparing with the accurate three-dimensional modeling method,the accuracy of average circle method by theoretical calculation was estimated and the maximum relative error of average circle method was less than 1.5%,which was in good agreement with the experimental results.Relative errors of the calculated and the experimental for obtaining the gear billet volumes with reference circle method are larger than those of the average circle method.It shows that average circle method possesses a higher calculation accuracy than reference circle method(traditional method),which should be worth popularizing widely in calculation of spur gear billet volume.
基金funded by Institutional Fund Projects under Grant No(IFPNC-001-611-2020).
文摘Landmines continue to pose an ongoing threat in various regions around the world,with countless buried landmines affecting numerous human lives.The detonation of these landmines results in thousands of casualties reported worldwide annually.Therefore,there is a pressing need to employ diverse landmine detection techniques for their removal.One effective approach for landmine detection is UAV(Unmanned Aerial Vehicle)based AirborneMagnetometry,which identifies magnetic anomalies in the local terrestrial magnetic field.It can generate a contour plot or heat map that visually represents the magnetic field strength.Despite the effectiveness of this approach,landmine removal remains a challenging and resource-intensive task,fraughtwith risks.Edge computing,on the other hand,can play a crucial role in critical drone monitoring applications like landmine detection.By processing data locally on a nearby edge server,edge computing can reduce communication latency and bandwidth requirements,allowing real-time analysis of magnetic field data.It enables faster decision-making and more efficient landmine detection,potentially saving lives and minimizing the risks involved in the process.Furthermore,edge computing can provide enhanced security and privacy by keeping sensitive data close to the source,reducing the chances of data exposure during transmission.This paper introduces the MAGnetometry Imaging based Classification System(MAGICS),a fully automated UAV-based system designed for landmine and buried object detection and localization.We have developed an efficient deep learning-based strategy for automatic image classification using magnetometry dataset traces.By simulating the proposal in various network scenarios,we have successfully detected landmine signatures present in themagnetometry images.The trained models exhibit significant performance improvements,achieving a maximum mean average precision value of 97.8%.
文摘In this paper, we discuss the precise asymptotics of moving-average process Xt =∞∑j=0 ajEt-j under some suitable conditions, where {εt, t∈ Z} is a sequence j=0 of stationary ALNQD random variables with mean zeros and finite variances.
文摘为解决传统目标检测算法在地铁、商场以及交通堵塞等地区高密度人群中因目标重叠和尺寸偏小而难以检测的问题,文中提出一种基于YOLOv5(You Only Look Once version 5)网络的目标检测算法。在算法模型的锚框部分引入新特征图来设计添加小目标检测层,以此提升检测小目标的准确性。通过重新定义一个卷积层,在YOLOv5中添加SOCA(Second-Order Channel Attention)注意力机制,提高了模型对复杂场景和遮挡的鲁棒性。引入Focal_EIoU(Focal and Efficient Intersection over Union)替换原始模型的损失函数CIoU(Complete Intersection over Union),从而提高了模型在高密度目标上的检测精度。实验结果表明,改进YOLOv5算法在CrowdHuman数据集上的平均检测精度比原模型提高了6.7百分点,召回率提高了3.8百分点,优于FPN(Feature Pyramid Network)和RetinaNet算法,实现了对高密度人群的目标检测优化。
文摘为解决在目标检测网络中使用特征融合方法带来的参数量大、计算复杂度高的问题,提出了一种融合无参注意力机制(SimAM)的特征融合方法。对动态蛇形卷积(DSConv)进行轻量化处理(Light-DSConv)。利用该结构自主学习目标几何形状的能力,对小目标的特征进行二次提取。利用SimAM模块对特征图空间域的重要性进行划分并与通道域权重相结合,进一步提升模型性能。在Pascal VOC 2007测试集上测试融合模块的有效性。结果表明:轻量化后,单个DSConv结构参数量下降85.6%。模型平均精度(mean average precision,mAP)比基线模型增加了4.41%,比添加现有特征融合方法模型平均增加3.78%。所提出模块的参数量、计算量、检测速度与现阶段其它方法相比均具有一定优势。
