支持语义的图像检索

如何跨越低层视觉特征和高层语义特征的鸿沟，是目前计算机视觉的一个难点。文中首先在低层视觉特征上提取图像的主要颜色，然后利用语义网络建立低层视觉特征和高层语义特征之间的关联，最后，为了提高检索效率，把相关反馈引进到图像检索系统中。实验证明。该方法可以应用到大量的数据库上，取得了较好的检索准确率。     

Waveform-based spaceborne GNSS-R wind speed observation: Demonstration and analysis using UK TechDemoSat-1 data

This paper explores two types of mathematical functions to fit single- and full-frequency waveform of spaceborne Global Navigation Satellite System-Reflectometry (GNSS-R), respectively. The metrics of the waveforms, such as the noise floor, peak magnitude, mid-point position of the leading edge, leading edge slope and trailing edge slope, can be derived from the parameters of the proposed models. Because the quality of the UK TDS-1 data is not at the level required by remote sensing mission, the waveforms buried in noise or from ice/land are removed by defining peak-to-mean ratio, cosine similarity of the waveform before wind speed are retrieved. The single-parameter retrieval models are developed by comparing the peak magnitude, leading edge slope and trailing edge slope derived from the parameters of the proposed models with in situ wind speed from the ASCAT scatterometer. To improve the retrieval accuracy, three types of multi-parameter observations based on the principle component analysis (PCA), minimum variance (MV) estimator and Back Propagation (BP) network are implemented. The results indicate that compared to the best results of the single-parameter observation, the approaches based on the principle component analysis and minimum variance could not significantly improve retrieval accuracy, however, the BP networks obtain improvement with the RMSE of 2.55 m/s and 2.53 m/s for single- and full-frequency waveform, respectively.     

INTERNET网检工具逻辑算符及检索表达式优化

对ＩＮＴＥＲＮＥＴ网检工具逻辑算符进行分析和比较。在此基础上，提出制定和优化检索表达式的方法。     

典型复合材料结构固化变形仿真预测与控制验证

针对典型复合材料结构固化成型过程中变形难以控制的问题,本文对典型复合材料结构的固化变形进行仿真预测,从固化工艺和模具补偿两方面对固化变形加以控制和验证。固化工艺方面以各设计点变形数据为基础确定了最优固化工艺曲线,模具补偿方面提出了一种构件有限元模型自适应调整的方法,综合考虑固化工艺参数与模具型面补偿采用了一种基于全局补偿量的协同控制方法。结果表明,通过仿真模拟L形构件的固化变形误差为12.4%,借助响应面优化算法得到的L形构件最优固化工艺曲线其固化变形预测值与各试验设计点最大变形的最小值偏差不超过3.3%;T形加筋壁板有限元模型经自适应调整后,对于下表面与目标型面之间的偏差距离,数值模拟值与试验测量值的最大相对误差为17.20%。通过全局补偿量的协同控制方法对半筒形壁板的模具进行补偿,其固化变形最大值相比于传统单一模具型面补偿控制方法降低了接近90%。     

支持快速数据检索的数据库机系统的研究

本文介绍了一个支持快速数据检索的数据库机TDM系统。TDM系统是一个树结构的关系数据库机系统。它具有紧密耦合式树型多机结构、数据的合理划分与分布、MIMD并行操作方式、多任务的合理调度与分配、多查询优化等特点,这使得数据库操作效率大为增加。该系统并行性高,功能齐备,十分适合于需要快速数据检索的应用场合。文中侧重介绍了TDM系统的功能、体系结构、系统构成及处理方式。     

网络环境下航空科技情报检索

介绍了常用的国内外航空科技文献数据库;阐述了航空科技情报的检索方法与检索途径;通过实例,说明了数据库检索与互联网信息检索结合使用的检索过程与检索技巧。     

Electron density retrieval from occulting GNSS signals using a gradient-aided inversion technique

In the coming years, opportunities for remote sensing of electron density in the Earth’s ionosphere will expand with the advent of Galileo, which will become part of the global navigation satellite system (GNSS). Methods for accurate electron density retrieval from radio occultation data continue to improve. We describe a new method of electron density retrieval using total electron content measurements obtained in low Earth orbit. This method can be applied to data from dual-frequency receivers tracking the GPS or Galileo transmitters. This simulation study demonstrates that the method significantly improves retrieval accuracy compared to the standard Abel inversion approach that assumes a spherically symmetric ionosphere. Our method incorporates horizontal gradient information available from global maps of Total Electron Content (TEC), which are available from the International GNSS Service (IGS) on a routine basis. The combination of ground and space measurements allows us to improve the accuracy of electron density profiles near the occultation tangent point in the E and F regions of the ionosphere.     

美家365网上商城检索系统的设计与实现

数据检索是大多数网站都具备的功能,不同的系统实现的方式千差万别,应根据实际需求找一个最优解决方案。本文从系统需求、系统分析、系统结构和流程、系统实现等几个方面详细阐述了美家365网站网上商城检索系统的设计与实现过程。系统不仅满足了该网站的业务需求,还对类似的系统开发有一定的参考价值。     

深度web资源搜索关键技术

深度web资源是指通过web搜索等形式能够访问的网络数据库资源,由于它与静态网页存在着一些差异,传统的搜索引擎技术不能够很好的实现深度web资源搜索.研究了深度web资源搜索中的若干关键技术,包括深度web资源的自动发现和基于本体的深度web数据抽取.通过实验验证了所提出技术的可用性与高效性.设计和实现了一种新的针对深度web资源的搜索引擎系统,能够获取深度web资源信息,并且利用这些信息抽取出结构化数据,从而为用户或者其他应用系统提供服务.已经应用在国家重大工程项目"国家科技基础条件平台门户应用系统"中,并取得了很好的应用效果.