几何语言与建筑设计

通过几何学中的点、线、面同造型艺术中点、线、面的联系，论述了几何形体对人的视觉心理感觉和形成的语言在建筑设计中的作用及重要性。并从中体现出建筑语言在人心中的反映。     

Coupling Magneto-Electro-Elastic node-based smoothed radial point interpolation method for free vibration and transient analysis of Functionally Graded Magneto-Electro-Elastic structures

A Coupling Magneto-Electro-Elastic (MEE) Node-based Smoothed Radial Point Interpolation Method (CM-NS-RPIM) was proposed to solve the free vibration and transient responses of Functionally Graded Magneto-Electro-Elastic (FGMEE) structures. By introducing the modified Newmark method, the displacement, electrical potential and magnetic potential of the structures under transient mechanical loading were obtained. Based on G space theory and the weakened weak (W²) formulation, the equations of the multi-physics coupling problems were derived. Using triangular background elements, the free vibration and transient responses of three numerical examples were studied. Results proved that CM-NS-RPIM performed better than the standard FEM by reducing the overly-stiff of structures. Moreover, CM-NS-RPIM could reduce the number of nodes while guaranteeing the accuracy. Besides, triangular elements could be generated automatically even for complex geometries. Therefore, the effectiveness and validity of CM-NS-RPIM were demonstrated, which were valuable for the design of intelligence devices, such as energy harvesters and sensors.     

基于径轴向Tchebichef矩函数三维曲面非对称结构的模型修正

采用积分的思想,将曲面机匣看作由若干圆柱机匣积累而成,将曲面机匣振型数据插值到一个圈数×点数×层数的三维像素空间内,能够实现径轴向Tchebichef矩(RAT矩)对曲面机匣模态振型的识别。利用RAT矩函数从曲面结构模态振型数据提取矩特征值,实现对振型数据的压缩,基于矩特征值的相关性分析能够很好地描述重模态现象,弥补了传统模态置信准则(MAC)值对于重模态描述的缺陷。进而将矩特征值代替振型数据,作为目标函数,应用于模型修正。矩特征值的数据量远小于振型数据,能够提高计算效率、避免修正结果不收敛。考虑到工程部件大多非完全轴对称,以一个带有非对称凸台的航空发动机机匣为例,基于RAT矩进行模态计算、相关分析和模型修正,发现同时利用频率与RAT矩进行模型修正,相关性有明显的提升,例如第9阶模态的相关性由35.35%提升到67.21%。各阶模态对的频差也都明显降低,最大频差由14.56%降至9.09%。证明了RAT矩函数能够应用在三维曲面非对称机匣的模型确认中。     

Adjustment of systematic errors in ALS data through surface matching

Surface matching is a well researched topic in both Computer Vision (CV) and terrestrial laser scanning (TLS) or ground based light detection and ranging (LiDAR), but the extent of the range images derived from these technologies is typically orders of magnitude smaller than those derived from airborne laser scanning (ALS), also known as airborne LiDAR. Iterative closest point (ICP) and its variants have been successfully used to align and register multiple overlapping views of the range images for CV and TLS applications. However, many challenges are encountered in applying the ICP approach to ALS data sets. In this paper, we address these issues, explore the possibility of automating the algorithm, and present a technique to adjust systematic discrepancies in overlapping strips, using geometrical attributes in a given terrain. In this method, the ALS point samples used in the algorithm are selected depending on their ability to constrain the relative movement between the overlapping laser strips. The points from overlapping strips are matched through modified point to plane based on the ICP method.     

基于神经网络的电表数字识别技术研究

提出了基于BP神经网络的机械式电表数字自动识别方法,首先通过预处理自动定位电表图像中的数字区域并实现单个数字的切分,然后对每个数字图像提取一组具有较高区分度且计算简单的典型网格特征,最后设计BP神经网络作为数字分类器,实现电度表显示值快速自动识别,该研究获得了电表数字正确识别率98.5%的结果,表明该系统具有较强的鲁棒性。     

广义均衡问题和不动点问题的收敛性定理及应用

运用平行算法，提出一种寻求广义均衡问题解及有限个严格伪压缩映像不动点集的公共元的新的迭代格式．在Hilbert空间中．证明了该算法的强收敛性。     

Hilbert空间上Lipschitz伪压缩映像族的强收敛定理

给出并证明了一个Lipschitz伪压缩映像族的强收敛算法,推广了Yao等人的结果。     

一种改进的模糊度子集选取方法及其性能评估

模糊度固定(ambiguity resolution, AR)能在一定程度上加快精密单点定位收敛并提高定位精度。但是模糊度固定率和计算效率会随模糊度数量的增加而降低,因此部分模糊度固定(partial ambiguity resolution, PAR)备受关注。PAR技术的关键在于模糊度子集选取,提出了一种改进的模糊度子集选取法,将高度角、信噪比以及模糊度方差联合作为模糊度子集选取的指标。实验结果表明：静态模式下,相比于全模糊度固定(full ambiguity resolution, FAR)和基于高度角的PAR方法,该方法的固定率分别提升9.65%和2.56%,首次固定时间分别缩短6.86%和3.43%,收敛时间分别缩短15.57%和5.13%,当三种方法固定率大致相同时,定位精度分别提升4.74%和5.39%;动态模式下,该方法的固定率分别提升22.75%和0.92%,首次固定时间分别缩短12.44%和0.44%,当三种方法固定率大致相同时,定位精度分别提升3.50%和4.89%。总体而言,无论是在静态还是动态模式下,该方法相比于FAR和基于高度角的PAR方法,性能均有所提升。     

基于原始观测值的UWB增强GNSS精密单点定位方法研究

全球卫星导航系统(global navigation satellite system, GNSS)可提供全球范围内全天候高精度导航、定位和授时服务。以精密单点定位(precise point positioning, PPP)为代表的绝对定位技术凭借定位精度高且全球一致、作业范围灵活等优势受到广泛关注,但是较长的收敛时间,限制了其在实时、快速精密定位应用中的使用。为解决上述问题,提出了超宽带(ultra-wideband, UWB)增强PPP方法,在多星座PPP中紧密集成UWB测距信息,以提高GNSS PPP性能。实验结果表明,在动态场景下,融合UWB量测使GPS/GAL双系统PPP在东、北、天3个方向的位置均方根(root mean square, RMS)值分别减少了76.99%、21.46%、64.53%,GPS/GAL/BDS三系统PPP减少了69.69%、37.21%、61.32%,并且收敛时间分别加快62.78%和57.75%。关于锚点数(几何构型)的评估表明,仅利用4个锚点就能将双系统和三系统3D误差RMS值减少67.98%、59.35%,收敛时间加快76.14%、62...     

基于改进YOLOv3网络的遥感目标快速检测方法

针对将来卫星在轨实时目标检测需求,且在其内存和算力都受限的条件下,提出一种改进的YOLOv3,利用轻量化网络代替YOLOv3的特征提取网络,实现遥感目标的高效检测。在目标检测精度相近的情况下,改进模型参数相比原先降低了1.5倍,计算量降低了3.3倍。同时提出了一种基于交并比的迭代聚类算法,分别在YOLOv3和改进YOLOv3上实现了7.0%和2.3%的平均精度均值(mAP)提升。实验表明:改进模型的检测速度最快能达到101 frame/s,当其mAP比YOLOv3高6%时,检测速度仍是YOLOv3的1.6倍。本文提出的改进YOLOv3是一种高效遥感目标检测方法,为未来星上应用打下基础。