目标检测是计算机视觉领域的研究热点和基础任务,其中基于锚点(Anchor)的目标检测已在众多领域得到广泛应用。当前锚点选取方法主要面临两个问题:基于特定数据集的先验取值尺寸固定、面对不同场景泛化能力弱。计算锚框的无监督K-means算...目标检测是计算机视觉领域的研究热点和基础任务,其中基于锚点(Anchor)的目标检测已在众多领域得到广泛应用。当前锚点选取方法主要面临两个问题:基于特定数据集的先验取值尺寸固定、面对不同场景泛化能力弱。计算锚框的无监督K-means算法,受初始值影响较大,对目标尺寸较单一的数据集聚类产生的锚点差异较小,无法充分体现网络多尺度输出的特点。针对上述问题,本文提出一种基于多尺度的目标检测锚点构造方法(multi-scale-anchor,MSA),将聚类产生的锚点根据数据集本身的特性进行尺度的缩放和拉伸,优化的锚点即保留原数据集的特点也体现了模型多尺度的优势。另外,本方法应用在训练的预处理阶段,不增加模型推理时间。最后,选取单阶段主流算法YOLO(You Only Look Once),在多个不同场景的红外或工业场景数据集上进行丰富的实验。结果表明,多尺度锚点优化方法MSA能显著提高小样本场景的检测精度。展开更多
Phrasal verbs,such as look after,look forward to,look through,etc.,account for a large per-centage in English vocabulary,and they are even more difficult than single words for.English learn-ers to learn.For one reason...Phrasal verbs,such as look after,look forward to,look through,etc.,account for a large per-centage in English vocabulary,and they are even more difficult than single words for.English learn-ers to learn.For one reason,they are of a large quantity by formation.By looking throughWebster’s Third New International Dictionary(1961),we can see how colorful the forms of展开更多
Previous reports indicated that the difference ofMHC-Ⅰ and Ⅱ between donor and recipient was criticalfor the occurrence of allogenic rejection or graft versushost reaction (GVHR). This study aimed at looking forthe ...Previous reports indicated that the difference ofMHC-Ⅰ and Ⅱ between donor and recipient was criticalfor the occurrence of allogenic rejection or graft versushost reaction (GVHR). This study aimed at looking forthe possibility of reducing the risk of immunologicalrejection through transplantation tolerance induced bytransduction of MHC gene. In the present study, wetransferred donor mice MHC class I gene (H-2k^b gene)展开更多
It was a week-end in summer and all the down trains were overcrowded. An old man was walking along the platform, looking for a vacant seat. Suddenly he saw one in a nonsmoker. The old man got in. A small bag was lying...It was a week-end in summer and all the down trains were overcrowded. An old man was walking along the platform, looking for a vacant seat. Suddenly he saw one in a nonsmoker. The old man got in. A small bag was lying on the seat and a well-dressed gentleman was sitting beside it. 'Is this seat vacant?' asked the old man.展开更多
To buffalo is to trick, to confuse,to intimidate, to frighten. To bulldoze also means to intimidate, to frighten.Long before the first Europeans arrived in the New world, a strangelooking animal lived on the
1.at the back of——in the back of——in back of——back of 四者同义,都表示“在……的后面”。at the back of多见于英国英语;后三者多见于美国英语,其中in back of和back of带有口语色彩。例: There is a garden at the back of ...1.at the back of——in the back of——in back of——back of 四者同义,都表示“在……的后面”。at the back of多见于英国英语;后三者多见于美国英语,其中in back of和back of带有口语色彩。例: There is a garden at the back of the house.展开更多
针对现有混凝土构件裂缝人工检测操作不仅费时、费力,而且易出现错检、误检、漏检,以及部分位置难以开展检测的问题,提出一种基于深度学习YOLOX(You Only Look Once)算法的混凝土构件裂缝智能化检测方法;首先采集、整理包含各类混凝土...针对现有混凝土构件裂缝人工检测操作不仅费时、费力,而且易出现错检、误检、漏检,以及部分位置难以开展检测的问题,提出一种基于深度学习YOLOX(You Only Look Once)算法的混凝土构件裂缝智能化检测方法;首先采集、整理包含各类混凝土构件的典型裂缝图像,并通过图像数据增强建立Pascal VOC数据集,然后基于Facebook公司开发的深度学习框架Pytorch,利用数据集训练YOLOX算法,并进行裂缝识别和验证;将训练完成后YOLOX算法移植至搭载安卓系统的手机端,进行现场实时检测操作。结果表明:在迭代次数为700时,混凝土构件裂缝识别精度可达88.84%,能有效筛分混凝土构件表面裂缝,并排除其他干扰项,证明了所提出的方法对裂缝具有较高的识别精度和广泛的适用性;经试验测试,移植至手机端的YOLOX算法能在提升便携性的同时保证高效、准确的检测效果,具有良好的应用前景。展开更多
远程塔台由于其低成本高时效远程实时控制技术正越来越受到民航业界的青睐,其中运动目标自动检测和显示是远程塔台的核心技术,作为增强现实技术更好地为管制员提供服务。在分析远程塔台机场场面背景复杂、场面目标多为远场景、小目标等...远程塔台由于其低成本高时效远程实时控制技术正越来越受到民航业界的青睐,其中运动目标自动检测和显示是远程塔台的核心技术,作为增强现实技术更好地为管制员提供服务。在分析远程塔台机场场面背景复杂、场面目标多为远场景、小目标等特点基础上,提出了一种改进的You Only Look Once(YOLO)算法来实现远程塔台运动目标的检测,算法核心思想以Darknet-53为基础网络,多尺度预测边界框,以运动目标图像坐标的偏移量作为边框长宽的线性变换来实现边框的回归,改善了传统YOLO算法损失函数不同大小的边框未做区分的问题,提高了检测准确性和速度。机场真实数据实验表明,该算法能快速、准确的检测出远程塔台的运动目标,并准确的回归运动目标边框及分类。展开更多
For the detection of marine ship objects in radar images, large-scale networks based on deep learning are difficult to be deployed on existing radar-equipped devices. This paper proposes a lightweight convolutional ne...For the detection of marine ship objects in radar images, large-scale networks based on deep learning are difficult to be deployed on existing radar-equipped devices. This paper proposes a lightweight convolutional neural network, LiraNet, which combines the idea of dense connections, residual connections and group convolution, including stem blocks and extractor modules.The designed stem block uses a series of small convolutions to extract the input image features, and the extractor network adopts the designed two-way dense connection module, which further reduces the network operation complexity. Mounting LiraNet on the object detection framework Darknet, this paper proposes Lira-you only look once(Lira-YOLO), a lightweight model for ship detection in radar images, which can easily be deployed on the mobile devices. Lira-YOLO's prediction module uses a two-layer YOLO prediction layer and adds a residual module for better feature delivery. At the same time, in order to fully verify the performance of the model, mini-RD, a lightweight distance Doppler domain radar images dataset, is constructed. Experiments show that the network complexity of Lira-YOLO is low, being only 2.980 Bflops, and the parameter quantity is smaller, which is only 4.3 MB. The mean average precision(mAP) indicators on the mini-RD and SAR ship detection dataset(SSDD) reach 83.21% and 85.46%, respectively,which is comparable to the tiny-YOLOv3. Lira-YOLO has achieved a good detection accuracy with less memory and computational cost.展开更多
With the increasing number of vehicles,manual security inspections are becoming more laborious at road checkpoints.To address it,a specialized Road Checkpoints Robot(RCRo)system is proposed,incorporated with enhanced ...With the increasing number of vehicles,manual security inspections are becoming more laborious at road checkpoints.To address it,a specialized Road Checkpoints Robot(RCRo)system is proposed,incorporated with enhanced You Only Look Once(YOLO)and a 6-degree-of-freedom(DOF)manipulator,for autonomous identity verification and vehicle inspection.The modified YOLO is characterized by large objects’sensitivity and faster detection speed,named“LF-YOLO”.The better sensitivity of large objects and the faster detection speed are achieved by means of the Dense module-based backbone network connecting two-scale detecting network,for object detection tasks,along with optimized anchor boxes and improved loss function.During the manipulator motion,Octree-aided motion control scheme is adopted for collision-free motion through Robot Operating System(ROS).The proposed LF-YOLO which utilizes continuous optimization strategy and residual technique provides a promising detector design,which has been found to be more effective during actual object detection,in terms of decreased average detection time by 68.25%and 60.60%,and increased average Intersection over Union(Io U)by 20.74%and6.79%compared to YOLOv3 and YOLOv4 through experiments.The comprehensive functional tests of RCRo system demonstrate the feasibility and competency of the multiple unmanned inspections in practice.展开更多
According to the infrared guidance ammunition(GA)attacking non-maneuvering targets on the ground or sea level,an improved bias proportional navigation(IBPN) is put forward,which can meet the constraints of the impact ...According to the infrared guidance ammunition(GA)attacking non-maneuvering targets on the ground or sea level,an improved bias proportional navigation(IBPN) is put forward,which can meet the constraints of the impact angle and the angle of attack(AOA). The motion equations and the collision triangle for the GA and the target are established in the two-dimensional plane. In accordance with the collision triangle, the integral value of the bias term is solved and BPN is designed on the basis of the relative velocity. To ensure the new method can be solved, closedloop equation of the IBPN is deduced. Considering the limitation of the AOA and the seeker angle, a four-phase IBPN is improved by setting different phases of the bias term. At the same time, the guidance law will make the impact angle achieve the desired angle and the normal acceleration also converges to zero. The simulation results show that the improved guidance law can be applied to various flight tasks and has great potential for engineering applications.展开更多
文摘目标检测是计算机视觉领域的研究热点和基础任务,其中基于锚点(Anchor)的目标检测已在众多领域得到广泛应用。当前锚点选取方法主要面临两个问题:基于特定数据集的先验取值尺寸固定、面对不同场景泛化能力弱。计算锚框的无监督K-means算法,受初始值影响较大,对目标尺寸较单一的数据集聚类产生的锚点差异较小,无法充分体现网络多尺度输出的特点。针对上述问题,本文提出一种基于多尺度的目标检测锚点构造方法(multi-scale-anchor,MSA),将聚类产生的锚点根据数据集本身的特性进行尺度的缩放和拉伸,优化的锚点即保留原数据集的特点也体现了模型多尺度的优势。另外,本方法应用在训练的预处理阶段,不增加模型推理时间。最后,选取单阶段主流算法YOLO(You Only Look Once),在多个不同场景的红外或工业场景数据集上进行丰富的实验。结果表明,多尺度锚点优化方法MSA能显著提高小样本场景的检测精度。
文摘Phrasal verbs,such as look after,look forward to,look through,etc.,account for a large per-centage in English vocabulary,and they are even more difficult than single words for.English learn-ers to learn.For one reason,they are of a large quantity by formation.By looking throughWebster’s Third New International Dictionary(1961),we can see how colorful the forms of
文摘Previous reports indicated that the difference ofMHC-Ⅰ and Ⅱ between donor and recipient was criticalfor the occurrence of allogenic rejection or graft versushost reaction (GVHR). This study aimed at looking forthe possibility of reducing the risk of immunologicalrejection through transplantation tolerance induced bytransduction of MHC gene. In the present study, wetransferred donor mice MHC class I gene (H-2k^b gene)
文摘It was a week-end in summer and all the down trains were overcrowded. An old man was walking along the platform, looking for a vacant seat. Suddenly he saw one in a nonsmoker. The old man got in. A small bag was lying on the seat and a well-dressed gentleman was sitting beside it. 'Is this seat vacant?' asked the old man.
文摘To buffalo is to trick, to confuse,to intimidate, to frighten. To bulldoze also means to intimidate, to frighten.Long before the first Europeans arrived in the New world, a strangelooking animal lived on the
文摘1.at the back of——in the back of——in back of——back of 四者同义,都表示“在……的后面”。at the back of多见于英国英语;后三者多见于美国英语,其中in back of和back of带有口语色彩。例: There is a garden at the back of the house.
文摘针对现有混凝土构件裂缝人工检测操作不仅费时、费力,而且易出现错检、误检、漏检,以及部分位置难以开展检测的问题,提出一种基于深度学习YOLOX(You Only Look Once)算法的混凝土构件裂缝智能化检测方法;首先采集、整理包含各类混凝土构件的典型裂缝图像,并通过图像数据增强建立Pascal VOC数据集,然后基于Facebook公司开发的深度学习框架Pytorch,利用数据集训练YOLOX算法,并进行裂缝识别和验证;将训练完成后YOLOX算法移植至搭载安卓系统的手机端,进行现场实时检测操作。结果表明:在迭代次数为700时,混凝土构件裂缝识别精度可达88.84%,能有效筛分混凝土构件表面裂缝,并排除其他干扰项,证明了所提出的方法对裂缝具有较高的识别精度和广泛的适用性;经试验测试,移植至手机端的YOLOX算法能在提升便携性的同时保证高效、准确的检测效果,具有良好的应用前景。
文摘远程塔台由于其低成本高时效远程实时控制技术正越来越受到民航业界的青睐,其中运动目标自动检测和显示是远程塔台的核心技术,作为增强现实技术更好地为管制员提供服务。在分析远程塔台机场场面背景复杂、场面目标多为远场景、小目标等特点基础上,提出了一种改进的You Only Look Once(YOLO)算法来实现远程塔台运动目标的检测,算法核心思想以Darknet-53为基础网络,多尺度预测边界框,以运动目标图像坐标的偏移量作为边框长宽的线性变换来实现边框的回归,改善了传统YOLO算法损失函数不同大小的边框未做区分的问题,提高了检测准确性和速度。机场真实数据实验表明,该算法能快速、准确的检测出远程塔台的运动目标,并准确的回归运动目标边框及分类。
基金supported by the Joint Fund of Equipment Pre-Research and Aerospace Science and Industry (6141B07090102)。
文摘For the detection of marine ship objects in radar images, large-scale networks based on deep learning are difficult to be deployed on existing radar-equipped devices. This paper proposes a lightweight convolutional neural network, LiraNet, which combines the idea of dense connections, residual connections and group convolution, including stem blocks and extractor modules.The designed stem block uses a series of small convolutions to extract the input image features, and the extractor network adopts the designed two-way dense connection module, which further reduces the network operation complexity. Mounting LiraNet on the object detection framework Darknet, this paper proposes Lira-you only look once(Lira-YOLO), a lightweight model for ship detection in radar images, which can easily be deployed on the mobile devices. Lira-YOLO's prediction module uses a two-layer YOLO prediction layer and adds a residual module for better feature delivery. At the same time, in order to fully verify the performance of the model, mini-RD, a lightweight distance Doppler domain radar images dataset, is constructed. Experiments show that the network complexity of Lira-YOLO is low, being only 2.980 Bflops, and the parameter quantity is smaller, which is only 4.3 MB. The mean average precision(mAP) indicators on the mini-RD and SAR ship detection dataset(SSDD) reach 83.21% and 85.46%, respectively,which is comparable to the tiny-YOLOv3. Lira-YOLO has achieved a good detection accuracy with less memory and computational cost.
基金supported by the National Key Research and Development Program of China(grant number:2017YFC0806503)。
文摘With the increasing number of vehicles,manual security inspections are becoming more laborious at road checkpoints.To address it,a specialized Road Checkpoints Robot(RCRo)system is proposed,incorporated with enhanced You Only Look Once(YOLO)and a 6-degree-of-freedom(DOF)manipulator,for autonomous identity verification and vehicle inspection.The modified YOLO is characterized by large objects’sensitivity and faster detection speed,named“LF-YOLO”.The better sensitivity of large objects and the faster detection speed are achieved by means of the Dense module-based backbone network connecting two-scale detecting network,for object detection tasks,along with optimized anchor boxes and improved loss function.During the manipulator motion,Octree-aided motion control scheme is adopted for collision-free motion through Robot Operating System(ROS).The proposed LF-YOLO which utilizes continuous optimization strategy and residual technique provides a promising detector design,which has been found to be more effective during actual object detection,in terms of decreased average detection time by 68.25%and 60.60%,and increased average Intersection over Union(Io U)by 20.74%and6.79%compared to YOLOv3 and YOLOv4 through experiments.The comprehensive functional tests of RCRo system demonstrate the feasibility and competency of the multiple unmanned inspections in practice.
基金supported by the China Postdoctoral Science Foundation(2013T60923)
文摘According to the infrared guidance ammunition(GA)attacking non-maneuvering targets on the ground or sea level,an improved bias proportional navigation(IBPN) is put forward,which can meet the constraints of the impact angle and the angle of attack(AOA). The motion equations and the collision triangle for the GA and the target are established in the two-dimensional plane. In accordance with the collision triangle, the integral value of the bias term is solved and BPN is designed on the basis of the relative velocity. To ensure the new method can be solved, closedloop equation of the IBPN is deduced. Considering the limitation of the AOA and the seeker angle, a four-phase IBPN is improved by setting different phases of the bias term. At the same time, the guidance law will make the impact angle achieve the desired angle and the normal acceleration also converges to zero. The simulation results show that the improved guidance law can be applied to various flight tasks and has great potential for engineering applications.
