视频监控中基于在线多核学习的目标再现识别

在非重叠多摄像机或单摄像机视频监控中,识别跟踪目标的再次出现很重要.针对传统支持向量机方法在特征融合方面的缺陷,本文提出了一种新的基于在线多核学习的人体目标再现识别方法.该方法对跟踪目标视频前景图像序列提取具有互补性的视觉单词树直方图和全局颜色直方图二种特征,再采用多核学习方法在线训练人体目标视觉外观,从而得到多核特征融合模型.实验结果表明,该方法能快速训练人体目标外观模型,满足视频监控的实时要求,多核融合模型获得了比单一特征模型和单核支持向量机方法更高的识别性能.     

基于复合分类特征的红外图像人体实时检测

针对红外图像中人体区域的检测问题,提出了一种基于复合分类特征的人体实时检测方法.首先根据红外图像中人体区域的特点,使用自适应的两级方向投影获得人体候选区域的可能位置,然后融合方向梯度直方图特征,人体形状特征及亮度分布惯性特征以充分描述人体区域的特点,并且采用支持向量机算法对候选目标中存在的人体进行分类检测.实验结果表明,本文提出的方法充分利用了复合分类特征各自的优点,具有较好的实时性和鲁棒性.     

基于直方图调整和击中击不中变换的弱小目标检测

针对复杂背景图像中弱小目标像素少、灰度低,难于检测的问题,本文提出了一种基于直方图调整和数学形态学击中击不中变换相结合的弱小目标检测方法.该方法根据弱小目标的灰度特征,先对原始图像进行直方图调整,提高弱小目标的邻域信杂比,然后对二值化图像进行击中击不中变换,检测出可疑目标集合.实验结果表明,该方法能够有效地检测复杂背景下的弱小目标.     

基于特征融合和交叉核SVM的快速行人检测方法

行人检测是目标识别领域的一大难点。现阶段用于行人检测的特征维数都比较高,为克服高维特征对实时性的影响,本文运用主元分析(PCA)对特征进行降维,加快检测速度。单一特征的信息有限,本文运用基于线性鉴别分析(LDA)的线性权重融合原则对一些底层特征(颜色、梯度、直方图)和多层次导向边缘能量特征进行特征融合使特征具有多源信息。且上述特征可采用积分图技术进行快速计算,所以行人检测系统的鲁棒性和实时性得到加强。在目标识别领域直方图交叉核支持向量机(HIKSVM)具有分类快,且准确率高的优点,采用其进行分类,系统实时性更进一步提升。实验表明本文方法检测速度和检测率优于经典的HOG+SVM算法。     

基于SVM的多生物特征融合识别算法

针对单生物特征识别的局限性,提出融合手背静脉和虹膜两种生物特征实现身份识别.基于尺度不变特征变换(SIFT)提取手背静脉的局部SIFT特征并对特征点进行匹配,利用特征匹配率作为手背静脉图像的相似度测度.通过Haar小波变换实现虹膜特征编码,利用加权汉明距对虹膜进行相似度测试.最后基于支持向量机(SVM)实现两种生物特征在匹配层的融合识别.利用CASIA虹膜数据库和TJU手背静脉数据库对算法性能进行测试,其等错率为0.02%,实验结果表明,该融合算法具有很高的识别性能,为生物特征识别研究提供了新思路.     

The Face Recognition Algorithm Based on Improved LBP

针对基本LBP算子提取的特征不够完整,不能全面地表达出人脸局部特征的问题,提出了基于分块的完备局部二值模式(CLBP)人脸识别算法.首先对原始人脸图像进行分块处理,对每一分块的图像进行局部差异值和中心像素灰度值分析,用Su2CLBP(8,2)、Mu2CLBP(8,2)和CCLBP(8,2)算子分别提取每一分块的直方图统计特征,然后将所有分块的CLBP直方图序列连接起来,得到人脸图像的CLBP特征,将其作为人脸的鉴别特征用于分类识别.最后利用Chi平方统计法计算直方图的不相似度,用最近邻准则进行分类.所提出的算法分别在ORL、FERET、YALE数据库中进行实验,分别取得了高达99.5％、92％、98.67％的识别率,与分块LBP算法相比识别率分别有2.5％、8％、2.67％的提高.实验结果表明,完备LBP提取的特征比较全面且具有较强的鉴别能力,在ORL.FERET、YALE人脸库中均能获得较好的识别率.     

多尺度加权LBP的人脸识别

针对传统局部二值模型(LBP)算子容易受到周围噪声点的干扰的缺点,提出了首先对图像进行Gauss滤波预处理,去除图像中的干扰噪声。针对传统LBP算子无法提取出非局部特征信息,提出一种新的基于多尺度加权的改进LBP(MWLBP)算子。MWLBP算子以不同大小的方型邻域为研究对象,将不同大小区域的LBP直方图进行加权求和。MWLBP比传统LBP算子提取的特征范围更大,在提取了局部特征的同时,保留了一定的非局部特征。相对于Gabor和其它特征提取方法,MWLBP算子在保留了多尺度特征的同时,能控制计算量大小。在ORL和Yale人脸数据库上的实验表明,Gauss滤波预处理确实能去除图像中的干扰噪声,提高识别准确率;MWLBP算子比传统的LBP算子、Gabor和其它特征提取方法减少了计算量,加快了分类器训练和人脸识别的速度,提高了准确率。     

基于改进的LBP人脸识别算法

针对基本LBP算子提取的特征不够完整,不能全面地表达出人脸局部特征的问题,提出了基于分块的完备局部二值模式(CLBP)人脸识别算法。首先对原始人脸图像进行分块处理,对每一分块的图像进行局部差异值和中心像素灰度值分析,用Su2CLBP(8,2)、Mu2CLBP(8,2)和CCLBP(8,2)算子分别提取每一分块的直方图统计特征。然后将所有分块的CLBP直方图序列连接起来,得到人脸图像的CLBP特征,将其作为人脸的鉴别特征用于分类识别。最后利用Chi平方统计法计算直方图的不相似度,用最近邻准则进行分类。所提出的算法分别在ORL、FERET、YALE数据库中进行实验,分别取得了高达99.5%、92%、98.67%的识别率,与分块LBP算法相比识别率分别有2.5%、8%、2.67%的提高。实验结果表明,完备LBP提取的特征比较全面且具有较强的鉴别能力,在ORL、FERET、YALE人脸库中均能获得较好的识别率。     

一种抽样小波域的遥感影像融合新方法

针对传统图像融合方法的不足,提出了一种新的基于IHS变换与抽样小波变换的高清晰遥感影像融合方法.该方法将直方图匹配后的全色影像的高频系数作为融合后的高频系数,而融合后的低频系数依据新提出的融合规则得到,最后采用小波逆变换与逆IHS变换获得融合图像.实验结果表明,该方法在提高融合影像的空间分辨率与光谱质量之间得到了更好的折衷.该方法的计算复杂度接近于常用的基于抽样小波变换的融合方法,明显少于基于非抽样小波变换的融合方法.     

结合边缘特征的遥感图像融合

曲波(Curvelet)变换是一种更适合于图像处理的多尺度几何分析(MGA)方法,具有很强的方向性.结合 HSI变换将其应用于全色图像和多光谱(MS)图像融合可以更好地表示图像中的有用特征.首先对多光谱图像进行 HSI变换,得到亮度分量 I,对全色图像和 I 分量进行曲波变换得到粗尺度系数和细节尺度系数,对全色图像的粗尺度系数和细节尺度系数进行叠加,计算归一化的全色曲波系数直方图,定义边缘有效因子,利用全色图像的特征信息对融合图像的粗尺度系数进行处理,对细节尺度系数采用函数对弱边缘进行增强,对新的曲波系数设计融合规则进行融合,逆变换后得到新的亮度分量 Inew,用 Inew 替代原亮度分量 I 进行逆 HSI 变换得到最终融合结果,采用统计类指标对融合结果进行评价.实验结果表明,该方法在保持光谱信息和提高空间分辨率上都有较好的效果.     

Insertion of triangulated surfaces into a meccano tetrahedral discretization by means of mesh refinement and optimization procedures

In this paper, we present a new method for inserting several triangulated surfaces into an existing tetrahedral mesh generated by the meccano method. The result is a conformal mesh where each inserted surface is approximated by a set of faces of the final tetrahedral mesh. First, the tetrahedral mesh is refined around the inserted surfaces to capture their geometric features. Second, each immersed surface is approximated by a set of faces from the tetrahedral mesh. Third, following a novel approach, the nodes of the approximated surfaces are mapped to the corresponding immersed surface. Fourth, we untangle and smooth the mesh by optimizing a regularized shape distortion measure for tetrahedral elements in which we move all the nodes of the mesh, restricting the movement of the edge and surface nodes along the corresponding entity they belong to. The refining process allows approximating the immersed surface for any initial meccano tetrahedral mesh. Moreover, the proposed projection method avoids computational expensive geometric projections. Finally, the applied simultaneous untangling and smoothing process delivers a high‐quality mesh and ensures that the immersed surfaces are interpolated. Several examples are presented to assess the properties of the proposed method.     

From Industry 4.0 to Research 4.0. Development of High Performance Fibers in the Light of Digitization

We associate a variety of innovations with the term ＂Industry 4.0＂. The pioneer of many 4.0 modifications forms the basisfor the trend towards the integrated di...     

The Research of Anti-Wear Properties of Organomolybdenum Compounds in the Engine Oils

A four-ball tester was used to evaluate the anti-wear performance of three kinds of organomolybdemun compounds in the engine oils, i. e., molybdenum dialkyldithiophosphate (MoDDP), molybdenum dialkyldithiocarbamate ( MoDTC), and sulphur and phosphorus freeorganomolybdeum (Molybdate). The results indicate that a low concentration of MoDDP doesn' t improve the anti-wear properties of the commercial engine oils, but a high concentration of MoDDP can obviously improve the anti-wear properties and the load-carrying capacity of the engine oils. MoDTC doesn' t improve the antiwear properties of the engine oils, but worsens the anti-wear properties of the oils. Signifi can timprove ment of frictional and wear characteristics is obtained with Molybdate added in the commercial engine oils and the formulated oils.     

Roman iron and steel: A review

ABSTRACT

The production of ferrous metal increased during the Roman Late Republican period, Principate and Empire. The direct bloomery process was used to extract the metal from its ores using slag-tapping and slag-pit furnaces. The fuel was charcoal and an air blast was introduced by bellows-operated tuyères. Iron formed as a bloom, often as a spongy mass of metal, which contained impurities from the smelting process, including unreacted ore, fuel, slag and fragments from the furnace walls, while the metal was often inhomogeneous with varied carbon contents. Blooms were either smithed directly into bars or ingots or they were broken up, which also allowed the removal of gross impurities and a selection of pieces with similar properties to be made. These could then be forge-welded together and formed into characteristically shaped ingots. Making steel in the furnace seems to have been achieved: it depended on the ore and the furnace and conditions within it. Surface carburization was also carried out. Iron and steel were used extensively in construction and for tools and weapons. Fire welding was often used to add pieces of steel to make the edges of tools and weapons, which could be heat-treated by quenching to harden them.     

Certification Institutions Expected to Update and Transfer Information on Present Standards

Standards are the basis for production enterprises to organize production, ex-factory inspection, trade (delivery) and technical exchanges, product certification, quality arbitration and supervision.……     

Flow calorimeter for enthalpy increment measurements on condensed gases

A flow calorimeter for enthalpy increment measurements on condensed gases is presented. A better knowledge of the properties of the liquefied natural gas is needed, and therefore a liquid loop has been designed for our flow calorimeter. The fluid loop in the calorimeter is designed in order to avoid the two-phase region, since two phases would give compositional disturbances in the measurements. The avoidance of the two-phase region is made possible by increasing the pressure of the test fluid after the measurement section, then heating the fluid at super-critical pressure past the critical point. Finally, the fluid is throttled to the low-pressure gas state at the inlet condition of the compressor that circulates the fluid. To perform the pressure increase, a new cryogenic pump has been designed. To evaluate the new equipment, measurements were taken on liquid ethane over the temperature range 146–256 K at pressure between 0.9 and 5.1 MPa.     

Synthesis of Alumina-Zirconia Nanocomposites by Solgel Process

Al₂O₃-ZrO₂ nanocomposites were developed starting with the solgel process. Composite alumina-zirconia nanopowders were synthesized from metallorganic precursors (Aluminium secondary butoxide and zirconium Iso propoxide) using the solgel process. The parameters affecting the synthesis—solvent, concentration of precursor, R/H ratio (i.e., dilution of water in solvent)—were varied as also the temperature and pH. BET and TEM were used to measure nanosize. Diffuse reflectance spectroscopy and also qualitative optical absorption led to identical particle size estimate. The variation of process parameters was used to study the effect and interdependence of process parameters. Artificial Neural Networks was used to rigorously analyze the process. Although this led to confirmation of interdependence of parameters, the presence of a single overwhelming solvent variable was also established. Then the optimal process was used to synthesize more nanopowder. To produce bulk nanocomposite the nanopowders were sintered by varying the temperature and time period. The sintered lithoids were probed with a vickers hardness tester to measure elastic modulus, hardness, and fracture toughness. The results showed high elastic modulus, modest hardness, and very high fracture toughness.     

Certification Becomes Important for the I Quality; and Safety Supervision of the Chinese Liquor Sector

The end of 2007, over 200 unit products of more than 80 Chinese firms have passed the quality grade certification for liquor products. These products involve distilled spirits, beer, wine, yellow wine, fruit syrup wine and others, and cover over 80% of the national top-branded liquor products.……     

Deposition and Coating Properties on CVD Tungsten

Surface characterization and microstructure studies are performed on chemical vapor deposited (CVD) tungsten coating. There is about 2 μm thickness diffusion layer of tungsten in the molybdenum substrate. The thermal shock test shows tungsten coating has good adhesion with molybdenum substrate, but the elements of oxygen and carbon in the tungsten coating have the bad affection to the adhesion. The result of high-temperature diffusion experiment is the diffusion rate from molybdenum substrate to tungsten coating is faster.