一种改进的从关系数据库挖掘本体的技术

通过对关系模式与OWL本体之间的相关性进行分析,提出了一种基于关系数据库自动生成本体的方法.该方法从关系模式主外键之间的相关性分析出发,对关系模式的结构和数据完整性约束等方面进行深入研究,并给出一系列从关系数据库挖掘本体的思路,能够挖掘深层次的语义信息,如继承关系和属性特性等.     

基于视觉推理的机器人多物体堆叠场景抓取方法

基于视觉的机器人抓取方法是研究智能机器人抓取问题的重要思路.本文提出了一种基于机器视觉推理的适用于多物体堆叠场景的机器人抓取方法,算法包含了场景理解和抓取规划两个步骤.在第一步的场景理解中,本文的算法包含两个主要部分:视觉操作关系推理和抓取部位检测.在视觉操作关系推理中,本文提出了一种基于深度卷积网络的视觉操作关系网络(Visual Manipulation Relationship Network, VMRN),以对物体和操作关系进行实时推理.在视觉操作关系网络中,通过设计物体对池化层,实现了物体检测和操作关系推理的端对端训练,提升了算法的速度和性能.在第二步的抓取部位检测中,本文提出了基于有向锚点框的全卷积视觉抓取部位检测网络,实现了对物体抓取部位的实时检测,并在康奈尔抓取数据集上取得了目前最高的精度.在抓取规划中,通过结合场景深度信息和抓取部位检测结果,获取当前被抓取物体的抓取点和对应的抓取向量,并通过坐标系变换将Kinect坐标系的抓取向量映射为机器人坐标系的抓取向量,完成当前抓取.实验结果表明,本文提出的机器人抓取方法能够在多物体堆叠环境下按照正确顺序完成抓取任务,并成功抓取目标物体.     

基于音乐基因的乐谱存储模型

本文提出一种基于音乐基因的乐谱存储模型S-MusicXML.将乐谱的存储和处理的基本单位由音阶提升到基因,有利于通过数据挖掘技术对音乐内涵的挖掘和存储.定义了旋律基因等概念,并通过实验进一步分析了挖掘音乐基因比挖掘音乐频繁模式更有优势.     

星敏感器星点聚类提取方法

星敏感器现有的星点提取方法运算时间长、存储需求大,从而限制星敏感器姿态初始捕获时间、轻小化等性能指标的提高.针对以上问题,提出了一种星敏感器星点聚类提取方法.首先,介绍和分析了星点聚类提取方法中的预处理滤波算法和聚类算法;然后,给出了该方法在北京控制工程研究所新研小型星敏感器中的实现方式和实现效果,该星敏感器在探月三期月地高速再入返回飞行器中成功实现了首次在轨飞行实验;最后,通过星敏感器观星实验对星点传统提取方法和星点聚类提取方法进行了比对验证,并通过小型星敏感器在轨飞行实验对星点聚类提取方法进行了在轨验证.实验显示,该方法提取星点所需时间最大约为传统方法的16%,星图存储需求不到传统方法的1%,且星点提取正确,在轨表现良好.结果表明,星敏感器星点聚类提取方法能有效减小星点提取所需时间,不需大容量星图存储器件,且逻辑简单、移植性好,适用于各类光学成像敏感器.     

餐厨垃圾自动称重系统的研究

针对目前餐厨垃圾收运工作存在无法自动计算污染者的餐厨垃圾产生量的问题,设计了一种以PIC单片机为核心的餐厨垃圾自动称重系统,通过安装在收运车辆上的RFID模块和自动称重模块,对不同污染者的垃圾产生量进行自动称重.数据通过I2C总线存储在IC卡中,作业完成提交IC卡,由计算机完成对污染者污染量的统计.测试结果证明,该系统能够对餐厨垃圾进行准确测量,达到自动称重的目的.     

基于知识点本体的语义扩展研究

传统的基于关键字的信息检索不能很好地检索到相关的资源.通过分析教学领域知识点本体的语义相关性特点,提出了一种基于知识点本体的语义扩展算法.该算法通过将概念间关系进行分离,利用概念层次结构和关联关系并考虑多种影响因素分别计算语义相似度和相关度,进而综合量化成语义关联度进行语义扩展检索.实验表明,本文所提出的语义扩展算法能够有效地提高查全率、查准率,取得较好的检索效果.     

GPU加速的表意式体模型线绘制算法

现有的体模型线绘制算法采用跟踪随机种子的方式得到特征线,存在严重的时空连续性问题.本文提出一种GPU加速的体模型线绘制算法,能够在每一帧抽取全部的特征线,从而彻底解决绘制中的不连续现象.我们首先提出一种并行的特征线抽取方法,在GPU的几何处理器中计算每个立方体单元中的特征线片段.为了解决可见性判定问题,我们采用splatting的方式生成深度图,取代已有方法中的光线投射方法,获得了明显的性能提升.采用自适应的深度偏移获得满意的可见性检测结果.为了将大量的体数据存储到高速显存中,我们通过剔除无关数据和压缩编码的方式,在不影响结果的前提下显著地减少了存储和带宽消耗.实验结果表明,本文算法能够生成时空连续的线绘制结果,并且绘制速度比已有CPU算法提高一个数量级.     

海量遥感数据存储管理技术综述

针对海量遥感数据的存储管理问题,以NASA EOS,World Wind,Google Earth,Google Maps,Bing Maps,Microsoft TerraServer,ESA,Earth Simulator,GeoEye,天地图,中国资源卫星应用中心,国家卫星气象中心,国家海洋应用中心等13个数据存储中心或系统为例,重点从遥感数据的存储组织方式与存储架构方面,综合与分析各个数据存储管理技术的特点与优势.通过综述国内外实际的数据存储与管理技术,有利于寻求更适合于海量遥感数据存储的技术与方法,应用于我国海量遥感数据存储管理的实际工作中.     

三掺杂化学计量比钽酸锂晶体的非挥发光存储性能

采用顶部籽晶助熔剂法生长了三掺杂近化学计量比Mg:Fe:Mn:Li Ta O3晶体,通过红外吸收光谱和居里温度研究了晶体缺陷结构.以蓝色激光为光源,获得了优异的光折变性质;采用多波长技术,研究了单晶的非挥发全息存储性能,得到了较高的固定衍射效率和灵敏度.蓝光具有较高能量,足以激发深(Mn)浅(Fe)陷阱中心的空穴,这大大提高了蓝光光折变性质和非挥发存储能力.在蓝色激光下,Mg2+不再是光损伤离子,而可以提高光折变特性.采用476 nm激光记录光栅,633 nm激光读取,在2.0 mol%Mg2+掺杂的晶体中获得了62.5%的固定衍射效率,非挥发全息存储的灵敏度提高到0.335 cm/J.     

结构与内容一体化搜索——从知识组织角度理解搜索

建立了基于医疗本体的知识库系统,通过结构和内容一体化搜索的方法,实现了对与搜索要求相关的信息的再组织,使搜索结果系统化、有序化、层次化,达到有效利用知识库中知识的目的.     

Effects of storage and processing on the nutritive value of certain Nigerian foods

The effects of storage and processing on the nutritive value of two Nigerian foods--raw cowpea (Vigna unguiculata) flour and palm oil, have been investigated. Increased retention of both thiamine and riboflavin as a result of increased water activity (Aw) were recorded for cowpea flour stored for 6 months. A storage temperature of 5 degrees C led to minimal losses of both vitamins in the stored flour. Differences in physico-chemical characteristics were observed between palm oil produced by the traditional method and that produced by a modern commercial method. Increasing Aw led to decreasing loss of unsaturation in the traditionally-produced palm oil during 4-week storage at ambient temperature (25 degrees C).     

Effects of storage and processing on the nutritive value of certain Nigerian foods

Summary The effects of storage and processing on the nutritive value of two Nigerian foods-raw cowpea (Vigna unguiculata) flour and palm oil, have been investigated. Increased retention of both thiamine and riboflavin as a result of increased water activity (A_w) were recorded for cowpea flour stored for 6 months. A storage temperature of 5°C led to minimal losses of both vitamins in the stored flour. Differences in physico-chemical characteristics were observed between palm oil produced by the traditional method and that produced by a modern commercial method. Increasing A_w led to decreasing loss of unsaturation in the traditionally-produced palm oil during 4-week storage at ambient temperature (25°C).     

Permeability of fresh and stored human erthrocytes to glycerol and its acylated derivatives

The effect of copper ions on the permeation of glycerol and its mono-, di-, and triacetate derivatives was studied in fresh and stored erythrocytes. Permeability was unchanged with storage. Acylated glycerols permeate the cells mainly by non-facilitated mechanism, as their transport is almost unaffected by copper ions.     

Triacylglycerol hydrolase: role in intracellular lipid metabolism

Recent scientific advances have revealed the identity of several enzymes involved in the synthesis, storage and catabolism of intracellular neutral lipid storage droplets. An enzyme that hydrolyzes stored triacylglycerol (TG), triacylglycerol hydrolase (TGH), was purified from porcine, human and murine liver microsomes. In rodents, TGH is highly expressed in liver as well as heart, kidney, small intestine and adipose tissues, while in humans TGH is mainly expressed in the liver, adipose and small intestine. TGH localizes to the endoplasmic reticulum and lipid droplets. The TGH genes are located within a cluster of carboxylesterase genes on human and mouse chromosomes 16 and 8, respectively. TGH hydrolyzes stored TG, and in the liver, the lipolytic products are made available for VLDL-TG synthesis. Inhibition of TGH activity also inhibits TG and apolipoprotein B secretion by primary hepatocytes. A role for TGH in basal TG lipolysis in adipocytes has been proposed. TGH expression and activity is both developmentally and hormonally regulated. A model for the function of TGH is presented and discussed with respect to tissue specific functions.     

露天矿库存管理系统中的数据库设计

在露天矿综合管理系统中，库存管理是一个重要环节，采购存储成本占矿山流动成本的15％～20％。所以露天矿物料及设备存储的管理优化是露天采矿优化中的重点。露天矿山仓储管理的现代化是实现矿山企业管理现代化的重要一环。在其数据库应用系统的设计中，数据库的结构设计是应用系统的基础。以库存管理系统的数据库设计为例，讨论了数据库的设计方法和技术，通过需求分析、概念结构设计、逻辑结构设计等方面阐述了库存管理系统数据库设计的全过程。     

Die Speicherung von Blutproteinen in den Histiozyten nach vorhergehender Histamineinwirkung

Summary By means of their special histological method, the storage of plasma proteins in the histiocytes was studied by the authors in rats and mice. If injected subcutaneously, both homologous and heterologous serum proteins were stored to the same extent at the site of the injection by the histiocytes of the connective tissue.After smearing the skin with histamine or after subcutaneous injection of histamine, intensive storage of protein occurs locally in the histiocytes. Because of the effect of histamine increasing permeability, these cells accumulate in their body the plasma proteins abundantly escaping from the blood stream.Even under physiological conditions, plasma proteins are constantly taken up and destroyed by the reticuloendothelial cells and this function has a fundamental biological importance.     

Permeability of fresh and stored human erythrocytes to glycerol and its acylated derivatives

Summary The effect of copper ions on the permeation of glycerol and its mono-, di-, and triacetate derivatives was studied in fresh and stored erythrocytes. Permeability was unchanged with storage. Acylated glycerols permeate the cells mainly by non-facilitated mechanism, as their transport is almost unaffected by copper ions.Portions of this work were included in the graduate thesis of Dr. Yaeger in partial fulfillment of the requirements for the M. D. degree.Acknowledgement. We thank Mrs Dalia Mazor for expert technical assistance.     

基于数据库缓存的Web应用性能优化解决方案

基于JAVA的应用在Web应用开发中占据着非常重要的地位。其中，许多系统选用Hibernate框架来负责系统数据的查询和存储，它能够在一定程度上提高系统的响应速度，优化系统性能。但在客户请求并发量大，数据库操作频繁的情况下，系统的性能还是无法保障。本文在Hibernate的基础上提出一种数据库缓存的解决方案来优化应用系统的性能。     

基于多本体的煤炭调度集成系统的研究

针对公司煤炭调度系统中总调度系统与下属分矿的子调度系统中存在异构数据源导致难以实现数据共享，设计了一个基于多本体的煤炭调度集成系统架构。重点研究了系统查询处理过程，首先根据全局本体到局部本体的映射，把针对全局本体的查询语句分解为对局部本体的子查询，再将子查询进一步重写为关系数据源的查询语句并给出了RDQL—to—SQL查询重写算法，最后把对关系数据源的SQL子查询结果进行集成返回给用户。