基于超像素的点追踪方法

目的由于当前大多数的追踪算法都是使用目标外观模型和特征进行目标的匹配,在长时间的目标追踪过程中,目标的尺度和形状均会发生变化,再加上计算机视觉误差,都会导致追踪的失误。提出一种高效的目标模型用于提高追踪的效率和成功率。方法采用分割后提取的目标特征来进行建模表示外观结构,利用图像分割的方法,将被追踪的目标区域分割成多个超像素块,结合SIFT特征,形成词汇本,并计算每个词在词汇本中的权值,作为目标的外观模型。利用外观模型确定目标对象的关键点位置后,通过使用金字塔Lucas-Kanade追踪器预测关键点在下一帧图像中的位置,并移动追踪窗口位置。结果结合点位移的加权计算有效地克服目标尺度和形状变化产生的问题。结论实验结果表明在目标发生形变或光照变化的情况下,算法也能准确地、实时地追踪到目标。     

Effects of a strong polyelectrolyte on the rheological properties of concentrated cementitious suspensions

Strong polyelectrolytes, referred to as superplasticizers, are known to improve the initial fluidity of concentrated cement suspensions. To quantify how the polyelectrolytes affect the fluidity, we have studied the effect of a strong anionic polyelectrolyte, melamine formaldehyde sulfonate (MFS), on the zeta potential of cement particles and on the steady-shear and low-amplitude rheological properties of cement suspensions. Adsorption of low concentrations of MFS onto the cement particles leads to an inversion in the sign of the surface potential, causing the electrostatically flocculated particles to become electrostatically dispersed and giving rise to a corresponding decrease in the steady-shear viscosity and storage modulus. At an intermediate MFS concentration, the steady-shear viscosity and the storage modulus each display a minimum. This concentration corresponds to that at which the zeta potential becomes constant. Larger concentrations of MFS result in an increase in the viscosity and storage modulus, which is attributed to depletion flocculation. These results thus relate the interaction between particles to the suspension fluidity through the analysis on the surface potential of particles and microstructure of suspension.     

基于VB和AutoCAD的纸盒辅助设计系统

在总结现有盒型特点并对其进行科学分类的基础上,结合Visual Basic语言与AutoCAD各自优势,开发具有盒型结构的参数化设计、材料使用率与消耗量、数据存储采集等功能的包装纸盒设计系统,达到提高绘图效率,简化开发过程,实现优化设计的目的。     

基于排队的服务系统的最小成本

用排队论的方法对基于排队的服务系统进行分析,确定了服务系统的排队模型.给出了统计平衡条件的排从系统的主要指标.讨论了该排队系统的最优化,并进行了实例分析.     

煤矸石混凝土无腹筋连续梁抗剪性能试验研究

对六根煤矸石混凝土无腹筋连续梁进行了抗剪性能试验,探讨了剪跨比对煤矸石混凝土无腹筋连续梁抗剪性能的影响及其变化规律,并与普通混凝土连续梁进行了对比分析,与《轻骨料混凝土结构设计规程》中的公式计算结果进行了比较。     

空气质量细粒度数据内联关系可视分析系统

随着数据观测和数字计算技术的快速发展,空气质量数据粒度越来越细致,为研究空气质量的细粒度数据内联关系可视化、辅助理解空气污染物的转化规律提供了数据基础.针对蕴含丰富内联特征的大气质量细粒度数据开发了一个内联关系可视分析系统.首先,基于大气质量数据多维属性设计时空数据维度模型增强数据的时空表达,并提出一种基于深度学习的特征提取方法,将空气质量数据从高维空间映射到支持交互可视分析的低维特征空间;然后,设计了一套多视图联动的可视化系统,帮助用户发现空气污染传输中的细粒度内联关系,理解污染传输途径的数据分布特征.用户实例和用户评价结果表明,该系统是有效的.     

基于深度学习的小样本声目标识别方法

声目标分类识别是声源识别领域的核心问题,然而,在应用深层神经网络进行声目标分类识别时,从少量样本中学习(样本复杂度较低)是一个具有挑战性的问题;针对此问题,提出了一种基于深度学习的小样本声目标识别方法,该方法将手工设计特征和对数梅尔声谱特征结合到一起,扩充了深度学习模型的可利用特征量,提高了声信号识别效率和精度;在实验验证中,该方法在测试集上实现了87.6%的识别精度;更进一步地,用较少量的训练样本对该方法和其它几种主流的深度学习模型的性能进行了比较验证,结果表明,该方法只需要更少量的数据即可实现同样的识别精度,在声源探测领域具有一定应用价值。     

Companion robots for older adults: Rodgers’ evolutionary concept analysis approach

Intelligent Service Robotics - This study aims to analyze the concept of companion robots for older adults from the perspective of nursing. This study employed a concept analysis. The literature...     

Growth and characterization of indium doped zinc oxide films sputtered from powder targets

Indium doped Zn O films were grown on quartz glass substrates by radio frequency magnetron sputtering from powder targets. Indium content in the targets varied from 1at% to 9at%. In doping on the structure, optical and electrical properties of Zn O thin films were studied. X-ray diffraction shows that all the films are hexagonal wurtzite with c-axis perpendicular to the substrates. There is a positive strain in the films and it increases with indium content. All the films show a high transmittance of 86% in the visible light region. Undoped Zn O thin film exhibits a high transmittance in the near infrared region. The transmittance of indium doped Zn O thin films decreases sharply in the near infrared region, and a cut-off wavelength can be found. The lowest resistivity of 4.3×10~(-4) Ω·cm and the highest carrier concentration of 1.86×10~(21) cm~(-3) can be obtained from Zn O thin films with an indium content of 5at% in the target.