地理信息技术在公安行业中的应用

分析警用地理信息系统的现状,总结建设中的关键GIS技术,并对系统的基本框架进行简要的描述,最后对具体业务应用进行简要的介绍。     

一种基于DOM的Web文档主题划分方法

主题划分是多主题文档自动摘要中的一个重要问题,提出了一种以网页结构为指导,利用页面对应DOM树中节点的自然分割功能以及相邻边界节点语义相似度的比较进行网页主题划分的方法.实验结果表明该方法具有较高的划分准确率,在此基础上抽取的网页摘要可显著增加文摘内容对原文的覆盖率、有效解决Web文档摘要分布不平衡问题.     

基于投影匹配的钢坯端面字符快速识别方法

由于高温及复杂光照的干扰等成像环境的影响,钢坯端面字符自动识别存在很多困难,其实时自动识别难度大。现存的字符识别方法难以适应实际生产现场的需求。文章提出了一种基于投影匹配的快速识别方法,通过将图象二值化后进行垂直投影,直接将一维投影结果进行模板匹配,降低运算维数。大量的现场实验证明,该方法可有效的减少识别时间,提高生产效率,字符的拒识率和误识率较低,可满足生产线的实时要求。     

大气湍流图像像素偏差机理分析

对大气湍流导致图像畸变和像素偏差的因素进行了深入分析,对大气湍流状态下激光发射远距离光束的光束漂移均方值进行仿真,分析光束漂移均方值曲线,为后续对图像畸变、像素偏差校正及克服湍流影响提供有效的理论依据。     

Casein kinase 2 down-regulation and activation by polybasic peptides are mediated by acidic residues in the 55-64 region of the beta-subunit. A study with calmodulin as phosphorylatable substrate

The noncatalytic beta-subunit is responsible for the latency of casein kinase 2 (CK2) activity toward calmodulin. Twenty-one mutants of the beta-subunit bearing either deletions or Ala substitutions for charged residues in the 5-6, 55-70, and 171-178 sequences have been assayed for their ability to substitute for wild-type beta-subunit as a suppressor of activity toward calmodulin. The only mutations that reduced the ability of the beta-subunit to suppress calmodulin phosphorylation activity, though being compatible with normal reconstitution of CK2 holoenzyme, were those affecting Asp55, Glu57 and the whole triplet Glu59-Asp-Glu61. The activity of CK2 holoenzyme, either native or reconstituted, toward calmodulin can be elicited by a variety of polybasic effectors, including polylysine, polyarginine, salmine, and histones H4, H3, and, to a lesser extent, H2a and H2b. Histone H1 and polyamines are conversely ineffective. The latent "calmodulin kinase" activity of CK2 can also be specifically unmasked by a peptide (alpha[66-86]) reproducing a basic insert of the catalytic subunit. This effect is reversed by equimolar addition of a peptide (beta[55-71]) including the 55-64 acidic stretch of the beta-subunit. Comparable polylysine stimulation was observed with the holoenzymes reconstituted with either beta wt or the beta mutants capable of assembling with the alpha-subunit, with the notable exception of those bearing Ala substitutions for acidic residues at positions 55, 57, and 59-61. These were nearly insensitive to 42 nM polylysine, which conversely promotes a more than 10-fold increase of calmodulin phosphorylation with wild-type beta.(ABSTRACT TRUNCATED AT 250 WORDS)     

Review of Broadband Metamaterial Absorbers: From Principles,Design Strategies,and Tunable Properties to Functional Applications

Metamaterial absorbers have been widely studied and continuously concerned owing to their excellent resonance features of ultra-thin thickness, light-weight, and high absorbance. Their applications, however, are typically restricted by the intrinsic dispersion of materials and strong resonant features of patterned arrays (mainly referring to narrow absorption bandwidth). It is, therefore essential to reassert the principles of building broadband metamaterial absorbers (BMAs). Herein, the research progress of BMAs from principles, design strategies, tunable properties to functional applications are comprehensively and deeply summarized. Physical principles behind broadband absorption are briefly discussed, typical design strategies in realizing broadband absorption are further emphasized, such as top-down lithography, bottom-up self-assembly, and emerging 3D printing technology. Diversified active components choices, including optical response, temperature response, electrical response, magnetic response, mechanical response, and multi-parameter responses, are reviewed in achieving dynamically tuned broadband absorption. Following this, the achievements of various interdisciplinary applications for BMAs in energy-harvesting, photodetectors, radar-IR dual stealth, bolometers, noise absorbing, imaging, and fabric wearable are summarized. Finally, the challenges and perspectives for future development of BMAs are discussed.     

Small Changes,Big Impact: Tungsten Bronzes with Extremely Large Second Harmonic Generation Achieved by the Transition Metal Doping on the B-Site

Two novel transition metal-doped tungsten bronze oxides, Pb_2.15Li_0.85Nb_4.85Ti_0.15O₁₅ (PLNT) and Pb_2.15Li_0.55Nb_4.85W_0.15O₁₅ (PLNW), are synthesized by high-temperature solid-state reactions. The Rietveld method using the high-resolution synchrotron radiation indicates that PLNT and PLNW crystallize in the orthorhombic polar noncentrosymmetric space group, Pmn2₁ (no. 31). As a class of tungsten bronze oxide, PLNT and PLNW retain a unique rigid framework composed of d⁰ transition metal cation (Ti⁴⁺ or W⁶⁺)-doped highly distorted NbO₆ octahedra along with the subsequently generated Pb/LiO₁₂ and PbO₁₅ polyhedra. Interestingly, the d⁰ transition metal-doped tungsten bronzes, PLNT and PLNW, exhibit extremely large second-harmonic generation (SHG) responses of 56 and 67 × KH₂PO₄, respectively. The observed immeasurably strong SHG is mainly attributed to a net polarization originating from the alignment of highly distorted NbO₆ octahedra with doped transition metals in the frameworks. It is believed that doping transition metal cations at the B-site of the tungsten bronze structures should be an innovative strategy to develop novel high-performance nonlinear optical materials.     

选择并融合粗细粒度特征的细粒度图像识别

目的 在细粒度图像识别任务中,类内图像姿势方差大,需要找到类内变化小的共性,决定该任务依赖于鉴别性局部的细粒度特征;类间图像局部差异小,需要找到类间更全面的不同,决定该任务还需要多样性局部的粗粒度特征。现有方法主要关注粗细粒度下的局部定位,没有考虑如何选择粗细粒度的特征及如何融合不同粒度的特征。为此,提出一种选择并融合粗细粒度特征的细粒度图像识别方法。方法 设计一个细粒度特征选择模块,通过空间选择和通道选择来突出局部的细粒度鉴别性特征;构建一个粗粒度特征选择模块,基于细粒度模块选择后的局部,挖掘各局部间的语义和位置关系,从而获得为细粒度局部提供补充信息的粗粒度多样性特征;融合这两个模块中提取到的细粒度特征和粗粒度特征,形成互补的粗细粒度表示,以提高细粒度图像识别方法的准确性。结果 在CUB-200-2011（caltech-UCSD birds-200-2011）、Stanford Cars和FGVC-Aircraft（fine-grained visual classification aircraft） 3个公开的标准数据集上进行广泛实验,结果表明,所提方法的识别准确率分别达到90.3%、95.6%和94.8%,明显优于目前主流的细粒度图像识别方法,相较于对比方法中的最好结果,准确率相对提升0.7%、0.5%和1.4%。结论 提出的方法能够提取粗粒度和细粒度两种类型的视觉特征,同时保证特征的鉴别性和多样性,使细粒度图像识别的结果更加精准。     

基于集成经验模态分解和最小二乘双支持向量回归机的风速预测算法研究

为有效解决风电大规模并网过程中面临的并网难和弃风等问题,实现可再生能源大规模平滑并网并保证大电网的安全稳定运行,采用集成经验模态分解(ensemble empirical mode decomposition,EEMD)和最小二乘双支持向量回归机(least square twin support vector regression,LSTSVR)算法进行风电场风速预测。分别介绍了LSTSVR、EEMD及自适应变异粒子群算法原理。给出基于EEMD和LSTSVR的风速预测流程,以安徽女儿岭风电场测风声雷达30、70 m处风速采样数据为例,开展基于EEMD和LSTSVR的风速预测算法验证,预测结果误差分析表明:基于EEMD+LSTSVR+自适应变异粒子群算法可以实现风电场风速的高精度预测。