Vision

摘要： In order to automatically detect the corrosion on the piston rod of hydraulic hoist of hydropower station and find the location of corrosion conveniently and accurately,a piston rod defect detection robot based on machine vision was developed.Firstly,the structure of the robot is presented,the detection system introduces the data collection process in the process of image processing,using a laser camera to capture image information.After the image preprocessing,cascade filter and morphological images after operation,the image information of gradient data,the depth of the defect data is extracted.To acquire the information of image defects in concave and convex,after using the RGB color image information acquisition defects the camera,HSV space separation and median filter and morphological processing are used to extracti defects in the image color information.Finally for defect information coding for quad,through comprehensive information for defect recognition and using QT development the human-computer interaction interface is developed,which can realize the reality and statistics of the robot′s motion control and defect information.Through the experimental verification of the robot′s performance,it is proved that the reliability of the robot′s motion and the accuracy of defect identification meet the design goal.

摘要： When firefighters are engaged in search and rescue missions inside a building at a risk of collapse,they have difficulty in field command and rescue because they can only simplymonitor the situation inside the building utilizing old building drawings or robots.To propose an efficient solution for fast search and rescue work of firefighters,this study investigates the generation of up-to-date digital maps for disaster sites by tracking the collapse situation,and identifying the information of obstacles which are risk factors,using an artificial intelligence algorithm based on low-cost robots.Our research separates the floor by using the mask regional convolutional neural network(R-CNN)algorithm,and determines whether the passage is collapsed or not.Then,in the case of a passage that can be searched,the floor pattern of the obstacles that exist on the floor that has not collapsed is analyzed,and obstacles are searched utilizing an image processing algorithm.Here,we can detect various unknown as well as known obstacles.Furthermore,the locations of obstacles can be estimated using the pixel values up to the bounding box of an existing detected obstacle.We conduct experiments using the public datasets collected by Carnegie Mellon university(CMU)and data collected by manipulating a low-cost robot equipped with a smartphone while roaming five buildings in a campus.The collected data have various floor patterns for objectivity and obstacles that are different from one another.Based on these data,the algorithm for detecting unknown obstacles of a verified study and estimating their sizes had an accuracy of 93%,and the algorithm for estimating the distance to obstacles had an error rate of 0.133.Through this process,we tracked collapsed passages and composed up-to-date digital maps for disaster sites that include the information of obstacles that interfere with the search and rescue work.

摘要： The fifth generation(5G)wireless communication networks are being deployed worldwide from 2020 and more capabilities are in the process of being standardized,such as mass connectivity,ultra-reliability,and guaranteed low latency.However,5G will not meet all requirements of the future in 2030 and beyond,and sixth generation(6G)wireless communication networks are expected to provide global coverage,enhanced spectral/energy/cost efficiency,better intelligence level and security,etc.To meet these requirements,6G networks will rely on new enabling technologies,i.e.,air interface and transmission technologies and novel network architecture,such as waveform design,multiple access,channel coding schemes,multi-antenna technologies,network slicing,cell-free architecture,and cloud/fog/edge computing.Our vision on 6G is that it will have four new paradigm shifts.First,to satisfy the requirement of global coverage,6G will not be limited to terrestrial communication networks,which will need to be complemented with non-terrestrial networks such as satellite and unmanned aerial vehicle(UAV)communication networks,thus achieving a space-airground-sea integrated communication network.Second,all spectra will be fully explored to further increase data rates and connection density,including the sub-6GHz,millimeter wave(mmWave),terahertz(THz),and optical frequency bands.Third,facing the big datasets generated by the use of extremely heterogeneous networks,diverse communication scenarios,large numbers of antennas,wide bandwidths,and new service requirements,6G networks will enable a new range of smart applications with the aid of artificial intelligence(AI)and big data technologies.Fourth,network security will have to be strengthened when developing 6G networks.This article provides a comprehensive survey of recent advances and future trends in these four aspects.Clearly,6G with additional technical requirements beyond those of 5G will enable faster and further communications to the extent that the boundary between physical and cyber worlds disappears.

摘要： For decades visual field defects were considered irreversible because it was thought that in the visual system the regeneration potential of the neuronal tissues is low.Nevertheless,there is always some potential for partial recovery of the visual field defect that can be achieved through induction of neuroplasticity.Neuroplasticity refers to the ability of the brain to change its own functional architecture by modulating synaptic efficacy.It is maintained throughout life and just as neurological rehabilitation can improve motor coordination,visual field defects in glaucoma,diabetic retinopathy or optic neuropathy can be improved by inducing neuroplasticity.In ophthalmology many new treatment paradigms have been tested that can induce neuroplastic changes,including non-invasive alternating current stimulation.Treatment with alternating current stimulation(e.g.,30 minutes,daily for 10 days using transorbital electrodes and^10 Hz)activates the entire retina and parts of the brain.Electroencephalography and functional magnetic resonance imaging studies revealed local activation of the visual cortex,global reorganization of functional brain networks,and enhanced blood flow,which together activate neurons and their networks.The future of low vision is optimistic because vision loss is indeed,partially reversible.

摘要： 精神实体的发现对奥古斯丁哲学具有奠基意义。根据《忏悔录》的叙述,奥古斯丁皈依前遭遇的核心困难,就在于无法摆脱物质或形体的(corporeal)空间范畴,无法将上帝和灵魂理解为真正的精神性存在。~①这一问题的最终解决得益于新柏拉图主义,后者的非形体观念使他能够突破受旧约、摩尼教和斯多亚主义深刻影响的形体主义,从而得以回到"内在自我"(inner self),并清楚确定地看到,无论自我还是自我之上的上帝都是精神性的存在。

摘要： Examining a post COVID-19 world order and humanity’s shared future The vulnerability of people across all nations in the face of the rampaging COVID-19 not only challenges the traditional conception of national security,but also highlights imperatives related to the vision of a community with a shared future for humanity。

摘要： 2018年年度旅游摄影师大赛(Travel Photog rapher of the Year,TPOTY)吸引了来自142个国家的摄影师,提交的参赛作品超过20000幅。除了万中选一的大奖作品,其他入围或提名的佳作通常较少进入大众视野,其实它们展现了更多精彩的''决定性一秒'',也为我们了解世界的多样性带来更多可能。

摘要： This paper aims to develop a platform that allows face features to be extracted faster using multiple algorithms for looking up people in a large database. We will be presenting an enhanced technique for human face recognition where we will be using an image-based approach (process of using two-dimensional images to create three-dimensional models) towards artificial intelligence by extracting features from face images by using Principle Component Analysis, Local Directional Pattern and SVM Machine Learning. Up until now, studies focusing on face recognition rely on the fusion of PCA (Principle Component Analysis) and LBP (Local Binary Pattern) for feature extraction, PCA and LBP were used for global feature extraction of the whole image and the features of the mouth area separately. Results show that this method was susceptible to random noise and resulted in a performance rate of 89.64% [1]. Also, recent studies have shown the fusion of PCA (Principle Component Analysis) and LDP (Local Directional Pattern) for feature extraction [2]. First, PCA is adopted to extract global features of facial images, then LDP operator is used to extract local texture features of eyes and mouth area and these areas are calculated by comparing the relative edge response value of a pixel in different directions. This fusion resulted in a performance rate of 91.61%. The results of PCA and LDP method show that it is more effective than adopting the fusion of PCA and LBP. It’s more robust to noise and improves the rate of facial recognition. However, both methods still suffer from changes in illumination, pose changes, random noise, and aging. In this paper, we propose using a set of trained images to make the facial recognition process faster and provide more accurate results.

摘要： 本发明涉及一种光路准直系统及方法，具体涉及一种基于TwinCAT Vision的二维镜架光路自动准直系统及方法，用于解决现有技术所存在的模块耦合性差、收敛速度慢的不足之处。该基于TwinCAT Vision的二维镜架光路自动准直系统，包括二维镜架、CMOS相机C1、CMOS相机C2、准直透镜L1、准直透镜L2以及TwinCAT控制系统，所述二维镜架包括反射镜M1和反射镜M2，以及对反射镜M1和反射镜M2进行姿态调整的两个步进电机。同时，本发明还提供一种基于TwinCAT Vision的二维镜架光路自动准直方法，该方法首先利用系数矩阵计算控制步进电机逼近目标坐标，再通过模糊控制规则逼近目标坐标，整个准直过程可在1秒内完成。

摘要： 本发明公开了基于Vision Transformer的滚动轴承故障智能识别方法，首先针对滚动轴承提取振动信号，进而将一维振动信号转化为二维灰度纹理图像，然后将其制作成滚动轴承故障诊断的样本，在此基础上设计一种Vision Transformer网络，利用该网络完成滚动轴承的振动灰度纹理图像分类，进而识别出滚动轴承的不同故障类型。本发明实现了基于Vision Transformer网络的滚动轴承故障智能识别，对保障旋转机械的运行稳定性和安全性具有重要意义。

摘要： 本发明提供一种基于Vision Transformer的跨域行人重识别方法，通过研究使用Vision Transformer的注意力网络用于视觉行人重识别网络，这是为数不多的不使用卷积神经网络结构用于图像匹配的行人重识别特征提取架构。其中，对于无标签的目标域数据，使用基于密度空间的聚类方法给数据赋予了伪标签。然后，将Vision Transformer特征提取架构与聚类模块融合，通过聚合信息存储模块对学习到聚合信息进行存储以学习全局特征。从而实现了面向跨域问题的基于Vison Tranformer的聚类方法。另外，还构建了样本dropout模块从而有效防止了训练过程趋向由这些硬样本引起的局部最小值的影响，为高效利用已标注源域和无标注目标域的行人数据训练一个泛化能力较强的网络模型。

摘要： 本发明提供了一种基于Vision框架和Core ML框架的物体追踪系统及方法，系统包括智能终端设备和与之蓝牙连接的云台设备，所述智能终端设备安装有APP，所述APP基于Vision框架和Core ML框架开发而成；通过智能终端设备处理器强大的处理能力，采集智能终端设备屏幕上的图像信息，计算出目标的位置坐标信息，算法转化后，发送指令控制云台设备转动，使追踪目标一值处于拍摄的中心，同时通过选定不同的识别模型，拓展可识别的范围，不再局限于人脸识别追踪，通过模型的不断训练、学习、升级，使识别准确率和效率不断的提升。

摘要： 本发明涉及一种基于Vision Transformer的图像压缩感知重构方法及系统，包括：(1)对图像依次进行预处理、分块压缩采样操作；(2)对图像通过训练好的初始重构网络进行初始重构，并进行重构和拼接；(3)对图像通过训练好的深度重构网络进行深度重构，得到深度重构图像。本发明提出了一种卷积神经网络和Vision Transformer结合的图像重构混合框架，是一种端到端的压缩重构图像方法，本发明使得图像压缩感知的网络结构能够继承卷积神经网络和Transformer的优点，并且最大程度保留全局和局部特征，提高图像重构的精度。

摘要： 本发明公开了一种基于轻量级Vision Transformer模块的故障诊断方法，包括如下步骤：获取多种不同故障类型的机械设备振动信号作数据集并进行训练集与测试集划分；利用二维图像编码方法将机械设备的一维振动信号编码为二维图像；利用轻量级Vision Transformer模块进行网络模型搭建并通过搭建的网络模型提取轴承数据集特征；利用转换好的轴承训练数据集对故障诊断模型进行训练；将测试数据集输入训练好的故障诊断模型得到预测的分类结果，根据分类结果判断待测机械设备的故障情况。所述方法具有特征提取能力强，诊断准确率高等优点。

摘要： 本发明涉及复杂环境下植物叶片病害检测和分类领域，具体为基于Vision Transformer模型的植物叶片病害检测分类方法，本方法以Vision Transformer模型进行特征提取和利用全连接层神经网络作为分类器对植物叶片病害识别分类。本发明通过采用新型Vision Transformer作为特征提取器，全连接网络作为分类器，搭配病害数据库，实现了以下目的：（1）对摄像机采集到原始图像，利用经典数字图像算法对原始图像经行预处理操作，提高送入网络图像数据的可检测性以及最大限度的简化数据；（2）使用新型视觉Transformer网络模型，针对高相似性和高复杂度的实际环境，实现较高的植物叶片病害检测准确度；（3）将检测结果和数据库内的数据经行比对分析，输出可能性最大的两个检测结果。

摘要： 本发明公开了一种基于Vision Transformer网络的弱监督实例分割方法、系统及介质，方法为：获取带标签的自然图像数据集和待分割自然图像；构建弱监督实例分割模型；所述弱监督实例分割模型包括ViT多标签分类模块及ViT候选区域评分模块；所述ViT多标签分类模块包括Vision Transformer网络及候选区域伪标签生成器；所述ViT候选区域评分模块包括候选区域生成器及ViT候选区域特征生成器；初始化弱监督实例分割模型，构建损失函数并在带标签的自然图像数据集上进行迭代训练，优化损失函数得到训练好的弱监督实例分割模型；将待分割自然图像输入训练好的弱监督实例分割模型中，得到实例分割结果。本发明实现了对自然图像的实例分割，在保持较高性能的同时，加快推理速度，减少算力的消耗。

摘要： 本申请公开一种GIGE vision数据传输方法、采集卡及系统，该方法包括：基于主机的下载命令接收来自于从设备的GIGE vision帧数据并拆包以获取数据包；对数据包进行拆分解析分流获得控制流和数据流；对分流后的数据流重组并缓存；根据分流后的控制流与主机通信并在主机处于待机状态时启动下载以将缓存的数据流传输至主机。用于克服相关技术中主机CPU和内存占用率，提高主机的处理性能。

摘要： 本发明提供一种基于Vision MLP的行人重识别方法，首先面向行人重识别任务，采用数据预处理将待测图像划分为预定维度及大小的的图像块序列。然后通过Vision MLP架构建立了行人重识别aMLP‑reID模型，设计了一个基于Vision MLP架构的线性注意力网络和MAL特征提取网络对全局依赖项建模用于特征提取，从而替代了vision transformer中自注意力层用于提取图像中深层特征，来进行行人重识别。而且由于采用线性注意力网络的结构减少了transformer架构中特征提取中平方量级的时间复杂度，因此本发明的基于VisionMLP的行人重识别方法也是为数不多的不使用卷积神经网络结构用于图像检索的行人重识别架构。