A global vision system: using hue thresholds to exact feature and recognize

The fast paced nature of robotic soccer necessitates real time sensing coupled with quick behaving and decision making. In the field with real robots, it is important to well perceive the location of ball, team robots and opponent robots through the vision system in real time. In this paper the architecture of global vision system of our small size robotic team and the process of object recognition is described. According to the study on color distribution in different color space and quantitative investigation, a method which uses H (Hue) thresholds as the major thresholds to feature exact and recognize object in real time is presented.

Global Vision

Innovative ZTE is a shining example of a Chinese company’s ongoing pursuit of excellence by Lu Qianwen EVERY autumn,28-year-old electronics fan Zhong Datong gets excited.It’s the time his favorite annual event,"P&T/ EXPO Comm China," the largest international information and communication technology exhibition in Asia,is held in Beijing. This year the dazzling new products at the booth of the ZTE Corp.were a contin-

Global Calibration Method for Monocular Multi-View System Using Rotational Dual-Target

To make the problems of existing high requirements of calibration tools, complex global calibration process addressed for monocular multi-view visual measurement system during measurement, in the paper, a global calibration method is proposed for the geometric properties of rotational correlation motion and the absolute orientation of the field of view without over lap. Firstly, a dual-camera system is constructed for photographing and collecting the rotating image sequence of two flat targets rigidly connected by a long rod at different positions, and based on the known parameters, such as, target feature image, world coordinates, camera internal parameters and so on, then the global PnP optimization method is used to solve the rotation axis and the reference point at different positions;Then, the absolute orientation matrix is constructed based on the parameters of rotation axis, reference point and connecting rod length obtained by this method. In the end, the singular value decomposition method is used to find the optimal rotation matrix, and then get the translation matrix. It’s shown based on simulation and actual tests that in comparison with the existing methods, the maximum attitude and pose errors is 0.0083˚ and 0.3657 mm, respectively, which improves the accuracy by 27.8% and 24.4%, respectively. The calibration device in this paper is simple, and there are no parallel, vertical and coplanar requirements between multiple rotating positions. At the same time, in view of the calibration accuracy, the accuracy requirements of most application scenarios can be met.

“Tsinghua Global Vision Lectures” Held

Since its launching on September 13, 2007, the Tsinghua Global Vision Lectures series, which has been incorporated into Tsinghua＇s elective course system, has received a warm welcome from students. In the past week, several famous university and company leaders delivered speeches in the series, including Professor Jeffery Lehman, Former President of Cornell University, Professor M.M. Salunkhe, Vice-Chancellor of Shivaji University in India, Mr. Maurice Greenberg, Former Chairman and CEO of American International Group （AIG）, Mr. Ronnie Chan, Chairman of the Hang Lung Group, and Mr. Joseph M. Tucci, Chairman, President and Chief Executive Officer of EMC Corporation.

University of Southern California Professor Kai Hwang Delivers Tsinghua Global Vision Lecture

University of Southern California （USC） Professor Kai Hwang delivered the Tsinghua Global Vision Lecture “Global Advances in Computer and Information Technology： Can China Take the Lead？” in the FIT Lecture Hall on January 6, 2009. During his lecture, Professor Hwang shared his ideas about several global advances in computer, networking, and information technology and discussed the great challenges yet to be confronted in information and technology in academia and in industry and the building of world-class universities. In his address,

Global Calibration Method of Multi-sensor Vision System Using Skew Laser Lines

Multi-sensor vision system plays an important role in the 3D measurement of large objects.However,due to the widely distribution of sensors,the problem of lacking common fields of view（FOV） arises frequently,which makes the global calibration of the vision system quite difficult.The primary existing solution relies on large-scale surveying equipments,which is ponderous and inconvenient for field calibrations.In this paper,a global calibration method of multi-sensor vision system is proposed and investigated.The proposed method utilizes pairs of skew laser lines,which are generated by a group of laser pointers,as the calibration objects.Each pair of skew laser lines provides a unique coordinate system in space which can be reconstructed in certain vision sensor＇s coordinates by using a planar pattern.Then the geometries of sensors are computed under rigid transformation constrains by taking coordinates of each skew lines pair as the intermediary.The method is applied on both visual cameras with synthetic data and a real two-camera vision system;results show the validity and good performance.The prime contribution of this paper is taking skew laser lines as the global calibration objects,which makes the method simple and flexible.The method need no expensive equipments and can be used in large-scale calibration.

我国体育高端智库建设的国际定位、现实困囿及推进策略

采用文献资料法、实地访谈法、对比分析法对我国体育高端智库的国际化定位、现实困囿与推进策略展开研究。研究表明,随着全球体育治理与体育事务国际化进程的加速推进,我国体育高端智库建设存在国际视野意识不强、国际战略研究程度不深、中国特色品牌尚未形成、成果国际传播不畅、国际议题设置实力较弱、智库人才的国际化培养不足等现实困囿。中国作为负责任大国应结合新时代体育全球化发展进程的现实需要,增强国际视野意识,深化国际战略研究程度,着力打造中国特色品牌,提高成果的国际传播能力,提升国际议题设置能力,加大体育智库国际化人才培养,为在国际上建设具有中国特色的体育高端智库提供理论和实践的战略支撑。

基于Pynq的半全局立体匹配算法实现

立体匹配算法作为双目立体视觉技术的核心算法,具有运算复杂度高、资源占用率高、实时性差的特点。而现有的基于FPGA实现方法多采用高性能FPGA芯片,利用其板卡丰富的资源进行纯verilog语言开发。这样做虽然实时性效果不错,但存在开发难度较大、周期较长、成本较高的弊端。论文针对该问题,提出了一种基于相对低廉的Pynq-z2开发板在PL+PS端协同开发的SGM立体匹配算法,将运算量大的部分在FPGA端进行硬件加速,部分代价优化算法在ARM端利用Python语言实现。实验表明,该算法可有效降低资源占用率,基本满足在相对低廉的嵌入式设备中实现的要求。

GNSS/SINS/视觉导航鲁棒算法

全球导航卫星系统(Global Navigation Satellite System, GNSS)、捷联惯性导航系统(Strapdown Inertial Navigation System, SINS)和视觉传感器优势互补,3者信息融合可获得高精度、无漂移的导航定位信息.针对GNSS/SINS/视觉融合导航易受运动速度、光照变化、遮挡等影响导致定位精度和鲁棒性降低问题,本文在图优化框架的代价函数中加入SoftLOne鲁棒核函数,设置量测值粗差检验程序,降低离群点带来的负面影响.进一步,对量测值计算残差进行卡方检验,对超限残差降权处理,提高系统精度和鲁棒性.实验结果表明,本文算法较不施加鲁棒核函数、不采用异常值剔除策略和卡方检验的传统算法,以及加入其他鲁棒核函数的算法精度更高、鲁棒性更好,能够较大程度提升GNSS/SINS/视觉导航定位精度和鲁棒性,在大尺度环境下,未出现较大漂移误差,绝对位姿均方根误差0.735 m,绝对位姿误差标准差0.336 m.

足球机器人全局视觉系统的快速标定方法

足球机器人全局视觉系统为决策系统提供机器人位姿信息,决策系统进行机器人的路径和动作规划;竞赛规则要求视觉系统必须现场安装调试,因此标定过程必须快速有效,且具有较高的精度;针对足球机器人全局视觉系统的特殊应用场景,提出了一种基于两步法简化标定模型的快速有效标定算法;利用近似欧拉角旋转矩阵建立摄像机标定模型,确定需要求取的摄像机内外参数,根据标定模型,对内外参数的求解过程进行了推导;借助场地上现有的一些标志点作为靶标,对模型的内外参数进行分步求取初始值,并利用L-M优化算法对摄像机内外参数进行分步优化求取精确值;通过实验分别从标定误差和机器人动作实现效率两个方面与其他标定算法进行对比,验证了该算法具有较高的标定精度;实验表明,本算法的优点在于不用借助专门的标定靶标,即可快速有效完成相机标定,标定精度可达±1 mm,完全能够满足决策系统的控制精度要求,且标定过程简单易实现。

基于实例分割的车间AGV精确定位研究

针对自动引导车(Automated Guided Vehicle,AGV)在复杂工业环境中定位误差大且准确率低的问题,本文提出了一种基于实例分割的车间AGV精确定位方法。通过SURF特征提取算法、最近邻匹配算法和随机一致性算法对双摄像头获取的图像进行实时拼接,使用动态加权融合算法消除接缝,利用RTMDet-Ins-x实例分割模型,获取AGV掩膜,结合AGV的特征进行定位,建立数学模型矫正定位误差。实验结果表明,在模拟车间环境下,该定位方法距离的最大误差为17.3 mm,平均误差为9.5 mm,角度最大误差为1.7°,平均误差为1.1°。与目标检测方法相比,该方法能够获取AGV的航向角度,距离误差缩小了64.4%,可实现车间AGV的精确定位,满足实际工程需要。

中国共产党“坚持胸怀天下”的辩证思考

坚持胸怀天下是中国共产党百年奋斗征程中形成的宝贵经验。中国共产党坚持胸怀天下始终贯穿着辩证思维,具体体现为:将坚持马克思主义与中华优秀传统文化相统一作为基因码,将坚持胸怀天下与心系祖国相统一作为视野观;将加强党的全面领导、全面加强党的建设与引领政党交流互鉴相统一作为引领力;将坚持全面推进与突出重点相统一作为方法论。

新时代中国共产党国际形象塑造研究的论域与展望

新时代新征程,中国共产党国际形象塑造研究对提升中国共产党的国际影响力具有重要的现实意义。纵览当前研究成果,学界从中国共产党国际形象塑造的历史进程、认同效果、价值意蕴、面临挑战以及实践路向等方面呈现了对该主题的研究论域。中国共产党的国际形象虽是一种“软”实力,但在推进中国式现代化的过程中发挥了显著的“硬”作用。展望未来,继续推进中国共产党的国际形象塑造的现代化实践,应当从深化理论建构、丰富研究方法、拓展研究视野等方面进一步提升新时代中国共产党国际形象塑造研究的学理性、科学性及针对性,增强中国共产党国际形象话语表达的说服力与影响力,努力向国际社会讲好中国共产党为人民谋幸福、为人类谋进步的故事。

激光雷达SLAM下移动机器人双目视觉全局定位

移动机器人在未知环境中执行任务时,需精准感知地图环境和其他物体位置,确定其自身位置。为了提高移动机器人的全局定位精度,提出激光雷达SLAM下移动机器人双目视觉全局定位方法。依据双目视觉系统测距原理,选取双目摄像机中的任意一个相机建立主坐标系,构建双目摄像机模型,结合棋盘图与张正友标定法,标定双目摄像机中参数。在此基础上,计算单帧图像特征点的尺度,完成双目视觉初始化,利用权重函数计算任意一个特征点与线、与面之间的加权值和距离值,估计移动机器人的位置,引入非迭代畸变补偿方法,完成对移动机器人的全局定位。测试结果表明,所提方法的定位结果与实际运动位置之间的最小误差为0.1 m,该方法可以精准定位到移动机器人位置,指导移动机器人更好地执行作业任务。

宽卷积局部特征扩展的Transformer网络故障诊断模型

视觉Transformer网络的高精度诊断性能依赖于充分的训练数据,利用卷积网络在提取局部特征上的优势,构造能同时描述故障局部和全局特征的提取层,提高诊断模型的抗噪声干扰能力。首先,引入卷积网络模块将原始振动信号转换为Transformer网络可以直接接收的特征向量,提取故障局部特征,并通过增加卷积网络的感受野。然后,结合Transformer网络多头自注意力机制生成的全局信息,构建能同时描述故障局部和全局特征的特征向量。最后,在Transformer网络的预测层,利用高效通道注意力机制对特征向量的贡献度进行自动筛选。在西储大学(CWRU)轴承数据集上的故障诊断结果表明,在信噪比-4 dB的噪声干扰下,改进后的Transformer网络轴承故障诊断模型的准确率达90.21%,与原始Transformer模型相比,准确率提高了13.2%,在噪声环境下表现出优异的诊断性能。

中国共产党坚持胸怀天下的逻辑探赜

坚持胸怀天下是党的百年奋斗历程的重要经验之一,是党的二十大提出的习近平新时代中国特色社会主义思想的世界观和方法论。从理论逻辑维度看,马克思人类解放思想和中华优秀传统文化为胸怀天下提供了理论支撑;从历史逻辑维度看,透视中国共产党胸怀天下的历史轨迹,彰显为世界谋发展的鲜明特质和理论品格;从实践逻辑维度看,中国共产党践行使命和重视党的自身建设铸就胸怀天下的独特优势,以人类命运共同体建设为契机谱写胸怀天下的时代篇章。基于历时态和共时态全景式立体化展现中国共产党胸怀天下的逻辑向度,可为世界谋发展、为人类谋进步提供重要启思。

基于自监督视觉Transformer的图像美学质量评价方法

现有的图像美学质量评价方法普遍使用卷积神经网络(CNN)提取图像特征,但受局部感受野机制的限制,CNN较难提取图像的全局特征,导致全局构图关系、全局色彩搭配等美学属性缺失。为解决该问题,提出基于自监督视觉Transformer(SSViT)模型的图像美学质量评价方法。利用自注意力机制建立图像局部块之间的长距离依赖关系,自适应地学习图像不同局部块之间的相关性,提取图像的全局特征,从而刻画图像的美学属性;同时,设计图像降质分类、图像美学质量排序和图像语义重构这3项美学质量感知任务,利用无标注的图像数据对视觉Transformer(ViT)进行自监督预训练,增强全局特征的表达能力。在AVA(Aesthetic Visual Assessment)数据集上的实验结果显示,SSViT模型在美学质量分类准确率、皮尔森线性相关系数(PLCC)和斯皮尔曼等级相关系数(SRCC)指标上分别达到83.28%、0.7634和0.7462。以上实验结果表明,SSViT模型具有较高的图像美学质量评价准确性。

基于视觉强化学习的数字芯片全局布局方法

在数字芯片后端设计中,全局布局需要同时兼顾线长与合法化,是一个组合优化问题。传统的退火算法或者遗传算法耗时且容易陷入局部最优,目前强化学习的解决方案也很少利用布局的整体视觉信息。为此,提出一种融合视觉信息的强化学习方法实现端到端的全局布局。在全局布局中,将电路网表信息映射为多个图像级特征,采用卷积神经网络(convolutional neural network,CNN)和图卷积网络(graph convolutional network,GCN)将图像特征和网表信息相融合,设计了一整套策略网络和价值网络,实现对全局布局的全面分析和优化。在ISPD2005基准电路上进行实验,结果证明设计的网络收敛速度加快7倍左右,布局线长减少10%~32%,重叠率为0%,可为数字芯片全局布局任务提供高效合理的方案。