Two-Staged Method for Ice Channel Identification Based on Image Segmentation and Corner Point Regression

Identification of the ice channel is the basic technology for developing intelligent ships in ice-covered waters,which is important to ensure the safety and economy of navigation.In the Arctic,merchant ships with low ice class often navigate in channels opened up by icebreakers.Navigation in the ice channel often depends on good maneuverability skills and abundant experience from the captain to a large extent.The ship may get stuck if steered into ice fields off the channel.Under this circumstance,it is very important to study how to identify the boundary lines of ice channels with a reliable method.In this paper,a two-staged ice channel identification method is developed based on image segmentation and corner point regression.The first stage employs the image segmentation method to extract channel regions.In the second stage,an intelligent corner regression network is proposed to extract the channel boundary lines from the channel region.A non-intelligent angle-based filtering and clustering method is proposed and compared with corner point regression network.The training and evaluation of the segmentation method and corner regression network are carried out on the synthetic and real ice channel dataset.The evaluation results show that the accuracy of the method using the corner point regression network in the second stage is achieved as high as 73.33%on the synthetic ice channel dataset and 70.66%on the real ice channel dataset,and the processing speed can reach up to 14.58frames per second.

ELLIPTIC EQUATIONS IN DIVERGENCE FORM WITH DISCONTINUOUS COEFFICIENTS IN DOMAINS WITH CORNERS

We study equations in divergence form with piecewise Cαcoefficients.The domains contain corners and the discontinuity surfaces are attached to the edges of the corners.We obtain piecewise C1,αestimates across the discontinuity surfaces and provide an example to illustrate the issue regarding the regularity at the corners.

Topological laser on square lattice with gain–loss-induced higher-order corner modes

We investigate the higher-order topological laser in the two-dimensional(2D) coupled-cavity array. By adding staggered on-site gain and loss to the 2D Hermitian array with a trivial phase, the system will emerge degenerate topological corner modes, which are protected by bulk band gap. For such a non-Hermitian model, by adjusting the parameters of the system and introducing the pumping into the cavity at the corner, a single-mode lasing with topological protection emerges.Furthermore, single-mode lasing exists over a wide range of pumping strengths. No matter where the cavity is initially stimulated, after enough time evolution, all the cavities belonging to the topological corner mode can emit a stable laser.

An angular blinking jamming method based on electronically controlled corner reflectors

Passive jamming is believed to have very good potential in countermeasure community.In this paper,a passive angular blinking jamming method based on electronically controlled corner reflectors is proposed.The amplitude of the incident wave can be modulated by switching the corner reflector between the penetration state and the reflection state,and the ensemble of multiple corner reflectors with towing rope can result in complex angle decoying effects.Dependency of the decoying effect on corner reflectors’radar cross section and positions are analyzed and simulated.Results show that the angle measured by a monopulse radar can be significantly interfered by this method while the automatic tracking is employed.

Topological edge and corner states of valley photonic crystals with zipper-like boundary conditions

We present a stable valley photonic crystal(VPC)unit cell with C_(3v)symmetric quasi-ring-shaped dielectric columns and realize its topological phase transition by breaking mirror symmetry.Based on this unit cell structure,topological edge states(TESs)and topological corner states(TCSs)are realized.We obtain a new type of wave transmission mode based on photonic crystal zipper-like boundaries and apply it to a beam splitter assembled from rectangular photonic crystals(PCs).The constructed beam splitter structure is compact and possesses frequency separation functions.In addition,we construct a box-shaped triangular PC structures with zipper-like boundaries and discover phenomena of TCSs in the corners,comparing its corner states with those formed by other boundaries.Based on this,we explore the regularities of the electric field patterns of TESs and TCSs,explain the connection between the characteristic frequencies and locality of TCSs,which helps better control photons and ensures low power consumption of the system.

The cornerstone of fiber-optic distributed vibration/acoustic sensing: Ф-OTDR

Fiber-optic distributed vibration/acoustic sensing(DVS/DAS) technology achieves breakthrough performance and explores broad cornerstone industrial applications.

Tailoring topological corner states in photonic crystals by near-and far-field coupling effects

We explore the behaviors of optically coupled topological corner states in supercell arrays composed of photonic crystal rods,where each supercell is a second-order topological insulator.Our findings indicate that the coupled corner states possess nondegenerate eigenfrequencies at theΓpoint,with coupled dipole corner states excited resonantly by incident plane waves and displaying a polarization-independent characteristic.The resonance properties of coupled dipole corner states can be effectively modulated via evanescently near-field coupling,while multipole decomposition shows that they are primarily dominated by electric quadrupole moment and magnetic dipole moment.Furthermore,we demonstrate that these coupled corner states can form surface lattice resonances driven by diffractively far-field coupling,leading to a dramatic increase in the quality factor.This work introduces more optical approaches to tailoring photonic topological states,and holds potential applications in mid-infrared topological micro-nano devices.

Closed corner divertor with B×▽B away from the divertor:a promising divertor scenario for tokamak power exhaust

A major challenge facing the steady-state operation of tokamak fusion reactors is to develop a viable divertor solution with order-of-magnitude increase in power handling capability as compared with present experience.A recently developed divertor concept for this end has been tested recently on EAST tokamak through combining the effects of a closed divertor corner and E×B drifts.The E×B drifts in the divertor move particles towards the outer divertor corner area in the scrape-off layer for B×▽B directed away from the divertor,which can significantly enhance the particle concentration there,facilitating divertor detachment.In recent EAST experiments,the effects have been demonstrated where the lowest electron temperature at the divertor plate is obtained with strike point located close to the corner in the horizontal target and with B×▽B away from the divertor.These experimental results are in reasonable agreement with SOLPS-ITER simulations including drift effects,suggesting that the new divertor concept potentially provides a promising divertor solution for long-pulse operations of future tokamak fusion reactors with much higher power fluxes.

DELIVERING THE GOODS China’s courier and logistics sector has thrived in its huge market and become a cornerstone of global supply chains

As China moves to fundamentally improve its virus response by downgrading COVID-19 management to Class B from top-level Class A,one sector that has remained crucial yet bruised by the three-year battle against the disease finally released a sigh of relief.

面向自动驾驶道路场景中异常案例的多模态数据挖掘算法

基于深度学习的视觉感知技术的发展有利于自动驾驶系统中环境感知技术的进步。然而,对于自动驾驶场景中的异常案例,目前的感知模型还存在一些问题。这是因为基于深度学习的感知模型的能力取决于训练数据集的分布。尤其是驾驶场景中的类别从未出现在训练集中,感知系统也往往很脆弱。因此识别未知类别和极端场景仍然是自动驾驶感知技术安全性的挑战。本文从处理数据集的角度出发,提出了一种新颖的多模态异常案例自动挖掘流程(corner case mining pipeline, CCMP)。为验证CCMP的有效性,在Waymo开放数据集的基础上构建了异常案例子集“Waymo-Anomaly”,该子集共有3 200个图像,每个图像都将包含文本中定义的异常案例场景。并且基于私有数据集Waymo-Anomaly,证明了CCMP针对异常案例场景挖掘的召回率可以达到91.7%。此外,还通过实验验证了目标检测器在包含异常案例的数据集中针对长尾分布的有效性。最终,希望从处理数据集的角度来提高自动驾驶感知模型在现实世界中的真实性。

CornerNet-Ghost:基于Hourglass-Ghost的轻量型目标检测模型

针对目前工业上目标检测任务较多,却限于设备原因无法流畅运行常规大型目标检测网络,对轻量型目标检测网络需求较大的问题,提出一种新型的轻量化目标检测模型:CornerNet-Ghost。采用特征提取网络Hourglass-Ghost作为骨干网络,对待测物体的左上和右下角点进行检测,并搭配级联角点池化优化提取的角点位置。实验结果表明,CornerNet-Ghost性能超过现有主流的轻量级角点检测网络CornerNet-Squeeze,且在检测计算时间远少于大型网络的条件下达到相近的准确性。

Passengers gathering-scattering analysis at the corners of subway stations

During both daily operation and emergency evacuation,the corners of walking facilities in subway stations play an important role in efficient circulation.However,the effectiveness of the corner is difficult to assess.In this paper,a method of passenger gathering and scattering analysis based on queueing models was proposed to investigate the corner performance in subway stations.Firstly,we constructed a set of state spaces of passenger flow according to passenger density and proposed the state transition model of passenger flow.Moreover,the model of passenger flow blocking and unblocking probability were also presented.Then,to illustrate the validity of the method and model,several passenger gathering-scattering scenarios and were simulated to verify the influence of passenger distribution and facility width on passenger walking,and the blocking probability,throughput,and expected time were also analyzed under various widths of the target corridor and arrival rates.Results showed that the proposed model can reproduce the trend of walking parameters changing and the self-organizing phenomenon of'faster is lower'.With the increase of arrival rates of passengers,walking speeds of passengers decrease and the expected walking time is prolonged,and the blocking probability sharply increased when the arrival rate exceeded 7 peds/s.In addition,with change of width of the target facility,efficiency of capacity of walking circulation facility fluctuated.With the width of the target corridor enlarged by 10%,the steady state of passenger flow was less crowded.Therefore,corridor width is critical to the circulation efficiency of passengers in subway stations.The conclusions will help to develop reasonable passenger flow control plans to ease the jam and keep passengers walking safely.

基于灰度均值的自适应FAST角点检测优化算法

光照不均、突变引起的灰度变化会影响图像特征检测效果。为此,设计一种基于灰度均值的自适应FAST-9-12角点检测算法。首先,利用特征点的延展性设计一种小面积双重检测模板,减少像素点与中心点的比较次数,提高区域正检率和检测速度;其次,依据图像局部灰度均值,在每个像素点检测模板内自适应调整阈值,避免灰度变化影响检测效果;最后,根据柔性非极大值抑制的思想设计角点半径抑制原则,筛选鲁棒性更强的角点。在Inria遥感影像数据集上的实验结果表明,FAST-9-12角点检测速度比FAST-12-16,FAST-9-16两种模板提高22%左右,因自适应阈值的提取方式不易受光照影响,检测准确率分别提高4.16和3.11个百分点,实现了图像特征快速和精准检测。

共同结构规范中舭部曲板作为硬角单元的适用条件研究

按照共同结构规范要求,对于船舶舭部板的船体梁极限强度和残存强度计算,规定在一般情况下应将其视为硬角单元,但未明确可以将其视为硬角单元的具体条件。本文利用共同结构规范对于曲板屈曲能力的评估方法和非线性屈曲有限元分析方法,对船舯区域和船舯以外区域的典型几何形状的舭部曲板进行了研究,得到了在共同结构规范要求的船体梁极限强度和残存强度计算中,舭部板可以被视为硬角单元的简要条件。

水下多格弹性角反射体声反射特性及强度增强方法

单格角反射体只能反射一定方向的入射波,为了反射不同方向入射声波,需要构造多格角反射体并研究其声反射特性。以不同类型八格金属板角反射体为对象,基于结构有限元耦合间接边界元法,对水下多格弹性角反射体的散射声场进行数值仿真,分析不同频率、材料、厚度、入射角度等因素影响下不同类型多格角反射体声反射性能,从便于工程应用角度提出增强其声反射能力的方法。通过典型情况下的水池实验,对仿真计算与实验数据的一致性进行验证。实验结果表明:水下多格弹性角反射体目标强度随入射频率增大而增大,并趋向稳定;目标强度随入射角度改变波动幅度较大,不同类型角反射体目标强度值差异较大;目标强度随平板厚度的增加而增大,但不显著;目标强度随着边长的增加而明显增大;软介质材料构成的多格角反射体,水下声反射性能明显优于金属角反射体。因此,为增强水下多格角反射体声反射能力,增大反射体边长及选用低阻抗材料是更有效的方法。

超高速公路自动驾驶车辆换道轨迹规划策略

为提高自动驾驶车辆在超高速公路行驶的安全性,提出了一种换道轨迹规划策略。首先,采用5次多项式生成一般变道轨迹簇,以车辆动力学极限和周围交通车辆为约束,将轨迹规划问题量化为求解换道行为持续时间;接着,考虑车辆动力学约束,建立了车辆动力学模型和Brush轮胎模型,基于所建立汽车模型的轮胎侧向力数据求解轮胎侧偏刚度,辅以魔术轮胎模型,验证所求轮胎侧偏刚度;然后,引入质心侧偏角-横摆角速度相平面,得到高速车辆安全驾驶包络线,并给定多组车速和附着系数进行CarSim仿真训练,确定满足车辆动力学约束的最短换道时间;最后,考虑与周围交通车辆的避撞约束,分析3种典型的换道场景,基于单障碍车的位置,确定满足避撞要求的最短与最长换道持续时间,建立满足安全换道要求的换道持续时间阈值模型。经多参数安全换道域检验,所建立的车辆安全换道持续时间边界模型能够在给定参数下求解出安全可行的换道轨迹,为超高速公路换道行为提供轨迹参考,提高超高速公路换道行为的安全性。

开挖宽度对异形基坑L形转角区支护结构变形影响研究

为研究不同开挖宽度对基坑L形转角区支护结构变形规律的影响,基于滨州市某异形深基坑工程项目的现场实测数据,分析了基坑开挖过程中L形转角区阴阳角位置处支护结构变形规律的差异性;建立了不同基坑宽深比(B/H,B和H分别为基坑宽度和深度)的L形转角基坑模型,研究了不同宽深比条件下转角位置处阴阳角相互作用条件下支护桩水平变形的空间分布规律,进一步明确基坑L形转角区支护桩不同深度处水平变形受影响范围。结果表明:当B/H<1.0时,阳角侧支护桩浅部水平变形受转角效应影响的范围随B/H的增加呈现出明显减小的趋势,其值近似满足2.4H-B;当B/H>1.0时,阳角侧支护桩浅部水平变形受转角效应影响的范围约为1.5H;坑内底部支护桩水平变形受基坑宽深比影响不明显,受转角效应影响的范围约为0.6H;当B/H<1.0时阴角侧支护桩浅部水平变形受影响范围约为1.6H,当B/H>1.0时其水平变形受影响范围约为0.6H+B;坑内开挖底面位置处阴角侧支护桩水平变形受影响范围约为0.6H+B,受宽深比变化影响不明显。

用平面六边形割角的细分曲面生成方法

平面六边形是一种更契合自然并具有独特优点的多边形,找到一种方便实用的平面六边形细分格式是细分曲面研究领域中的基础工作之一。本文提出了一种平面六边形细分方法,对任意初始凸多面体网格,只用平面六边形进行割角细分,可以使得细分曲面光滑、保凸、具有插值性质,且细分过程中新增的面片都是平面六边形.我们将该方法简称为平面蜂窝细分.论文给出了平面蜂窝细分的拟线性四点格式几何规则及其细分矩阵,其几何意义直观,相应的算法简单、可行且数值稳定.论文分析了细分曲面的收敛性和光滑性,给出了C光滑性条件及其证明.为了提高细分曲面的光滑性,传统的方法是使边数不同的多边形面片按等比例收缩,该方法对于平面蜂窝细分没有预期效果并且无法处理三角形特殊面;论文给出了一种新方法,获得的割角参数使得细分曲面更加光滑,且可以统一处理三角形、四边形等特殊面,从而避免特殊点/面带来的局限性.论文还提出了平面蜂窝细分方法的一种推广和一种自然边界处理方法,并讨论了平直边界和退化情形下尖锐特征生成的方式,可用于细分任意可定向初始网格.文中给出了一些细分曲面的例子,并与经典的细分方法进行了比较,验证了新方法的有效性和具有的优点.

融合角点检测的点云配准研究

为提高点云配准效率,提出一种基于点云曲率信息的Harris角点提取算法,通过引入曲率约束,对传统Harris角点提取做出改进;计算每个点的曲率值并与整幅点云的平均曲率作比较,保留曲率值大于平均曲率的点云,实现对配准价值较低的点的剔除,为后续点云配准减少计算量;构建所提取角点之间的特征匹配,结合采样一致性(sample consensus initial alignment, SAC-IA)粗配准和正态分布变换(normal distributions transform, NDT)精配准,寻找最优的变换矩阵,实现源点云和目标点云的重合。通过比较不同算法在公开数据集上进行实验验证,结果表明该方法对复杂点云、多拐点点云都有较好的表现,对初始位置相差较大的点云也有良好的适用性。

压力敏感涂料测量技术在掠型叶栅表面测压中的应用

为了获得某掠型叶片吸力面压力分布,对一稠度为1.53,掠角为10.7°的掠叶栅进行了光路布局以及压力敏感涂料(PSP)压力测量,利用图像处理和三维重构算法得到了可读性高、便于数据提取的三维叶片吸力面压力分布,并和测压孔测量结果进行了对比。结果表明:利用自主发展的光路布局方法,有效解决了叶片表面光照辐照度不均匀、拍摄视角受限的问题;得到了高信噪比的原始压敏图像以及宽范围多工况的掠叶栅吸力面全域压力分布。与测压孔测量相比,PSP测量误差在5%以内,并且PSP测量结果还捕捉到了峰值等熵马赫数位置沿弦向的迁移以及叶片尾缘的气流分离现象。对不同攻角和来流马赫数下的掠型叶片吸力面压力分布进行分析:掠型叶片在零攻角和正攻角下存在明显的径向压力梯度,促使后掠型叶片气流向前掠迁移,增加了前掠下流低能流体的能量,使得前掠更有利于延迟角区分离,后掠则更容易发生角区分离;掠型叶片在低来流马赫数下展向压力对称性良好,掠设计的影响不显著,高来流马赫数则放大了前后掠对角区分离控制效果的不同。相比传统测压孔有限的压力分辨率,PSP测量可为掠型叶片设计和流动机理分析提供丰富可靠的压力实验数据。