面向地学应用的遥感影像分割评价

影像分割是面向地理对象影像分析(GEOBIA)中的一个关键环节。分割评价有助于为影像选择合适的分割方法和最佳分割尺度。本文提出一种以遥感地学应用面向的对象为依据的分割方法分类及评价体系。首先将影像分割方法分为面向局部特征监测的典型目标识别和面向全局特征监测的面向GEOBIA的分割方法两组,进而针对这两组分割方法提出了两套分割评价测度指标及相应的综合评价方法。在面向典型目标识别的分割方法评价中,使用区域内部非均质度、区域间灰度对比度、区域间散度对比度、边界点梯度和单位像素运行时间作为评价测度,并针对由于评价测度间的相关性而无法直接确定各测度权重分配的问题,提出利用熵权法为各个评价测度分配权重以获得综合评价结果的分割评价方法,该评价方法可用于选择合适的分割方法。用于面向GOEBIA的分割方法中使用分割区域内均质性和区域间异质性作为评价测度,这种评价方法适用于选择最优分割尺度参数。本文通过定量试验论证了这两种评价方法的有效性,试验结果表明其在遥感应用中具有实际意义。最后本文分析了影像分割评价方法的不足及未来的发展方向。     

天文学英语新词自动提取系统

科技名词中文译名标准化和规范化工作是推动我国科技进步和科学知识传播的重要基础。新的科学概念和技术名词层出不穷,如何及时发现并确定新生术语的中文译名是一项普遍的社会需求。介绍了一套全新的天文学英语新词自动提取系统。该系统综合使用脚本过滤、术语识别、正则表达匹配等多种方法,能够自动追踪Ar Xiv论文数据库的更新,分析天文学论文的内容,生成推荐术语列表,从而将学科专家从繁重的科技新词收集整理工作中解放出来,把有限的精力集中到更能体现专业素养的新词审定工作当中。这个系统将为推动天文学等基础学科的新词收集,乃至学科标准化等工作发挥积极作用。     

基于Fisher判别分析法的雹云识别模式及其效果检验

作者根据雹云等强对流天气的雷达回波参数的数量特征，应用Ｆｉｓｈｅｒ判别分析法，对待识别的强对流天气样本云作出雹云和雷雨云的定性分类判别。用该方法对成都市１９８２～１９９０年的１７３个雹云和雷雨云样本回波资料建模，并对１９９０年的１３个试报样本进行检验。其判别结果与主分量分析法判别结果比较表明该方法用于雹云识别是适合的。     

地图数字化的数学基础及在世界地图数字化中的应用

地图作为地图数据库、地理信息系统、遥感等最基础和最重要的数据源 ,其精度问题已成为各学科共同关注的问题。本文提出了地图数字化的数学基础 ,详细讨论了两种常用的数据采集模型 :并以世界地图数字化系统为实例 ,介绍了系统的构成 ;并重点讨论了破译地图数学基础的原理和方法。     

一个动物识别专家系统在PROLOG下的实现

首先扼要分析了专家系统的基本工作原理，然后用PROLOG语言实现了一个简单的动物识别专家系统。     

双链DNA片段的碱基空间模式BSP——基因组调控区的一种可能的信息编码方式

生物体可看作是一种特殊的信息处理系统 ,论文的第二章从这种观点出发讨论了生物细胞内最大的信息源基因组DNA序列的信息编码问题。首先探讨了基因组蛋白质编码区的信息编码问题 ;简要回顾转录水平的基因表达调控后 ,讨论了基因组基因表达调控区的信息编码问题。第三章介绍了我们利用internet上可获取的开放的生物信息学资源 ,对基因组DNA调控区的一种可能的编码方式所进行的理论研究。这种可能的编码方式我们称之为双链DNA上的碱基空间模式BSP (BaseSpatialPattern)。首先 ,我们发现了由三联体BSP介导的一种 4与 2 0的关系 ,称为 4 -BSP -2 0关系。并推测 4 -BSP -2 0关系也许是某些DNA—蛋白质相互作用中的一种识别密码 ;第二 ,研究真核细胞转录调控中的DNA—蛋白质相互作用的实验数据表明 :某些DNA结合蛋白质对双螺旋核酸的专一性识别可能与BSP有关。并找到了 7组由其他生物学家公开发表的实验数据支持这一猜测 ;第三 ,研究了BSP的形式化描述问题 ,找到了一种基于等价类的BSP描述方法 ,并开发了相应的计算机处理算法。该描述方法简明且便于用计算机进行处理 ,更重要的是它提供了一种研究途径 ,可把BSP和 4 -BSP -2 0关系推广到单链DNA (ssDNA)和ssDNA -蛋白质相互作用的研究中 ,从而可进一步推广到RN     

AN AUTOMATIC PROCEDURE FOR ESTIMATING SPRING SMALL GRAINS ACREAGE FROM LANDSAT DATA

The development and testing of a computerized technique for utilizing multitemporal Landsat multispectral scanner data to estimate spring small grains acreage are described. The Kauth-Thomas green number transformation is applied to the data prior to application of the pixel-level standardized logic. Four years of data from the U.S. and Canada are used in testing. A comparison of the test results to those of the analyst-intensive procedures indicates a potential for increase in near-harvest estimation efficiency with little sacrifice in accuracy.     

Spatial pattern recognition by decomposition

The decomposition problem—the assignment of sample observations to component populations—is studied in a spatial context. The observations are spatially located and the assignment to component populations takes into account the value of each observation as well as the values of neighboring observations. Both parameter estimation and assignment rules use a new method that integrates a standard multivariate decomposition algorithm with nonlinear regression. The method is illustrated and tested with artificial data. The distribution of the trace component Cr₂O₃ in recent Lake Michigan sediments is, then, analyzed by the method. It yields a pattern of component populations that is correlated with the Lake's bottom structure and depositional environments. Study carried out as visiting scholar at the Department of Geological Sciences, Northwestern University, Evanston, Illinois U.S.A.     

APPLICATION OF PATTERN RECOGNITION TO SPECIATION DATA OF HEAVY METALS IN SUSPENDED PARTICULATES OF URBAN AIR

In order to get a better knowledge of the air pollution levels of the urban air of Barcelona,twelvesamples from three stations of the air-monitoring control network were analysed.Each control station isrepresentative of a different environmental situation.Thus the station at Plaa Molina responds to asimple domestic and urban traffic situation;the station at Poble Nou to the influence of medium andsmall industries,mainly metallurgical;and the station at Montcada to a zone in which there existimportant industries,a cement factory,two highways and railroads.The total content of unsedimentable particles,three inorganic anions(chlorides,sulphates andnitrates)and six heavy metals(Fe,Pb,Cu,Mn,Cr and Cd)were determined on all the samples obtainedduring the same period of time.The use of an analytical scheme of sequential extraction has alsopermitted the subdivision of the total content of each metal into the following four fractions:(1)solubleor exchangeable;(2)associated with carbonates and/or with iron and manganese oxides;(3)bound toorganic matter;and(4)residual.The use of principal component analysis applied to the data as a whole showed the associations andcorrelations between the various parameters as well as between the different above mentioned fractions.The predominant presence of Cd in the‘soluble or exchangeable’fraction is confirmed,while the othermetals are mainly associated with the‘carbonates and/or Fe and Mn oxides’and‘bound to organicmatter’fractions.Hierarchical cluster analysis applied to the twelve samples showed clearly the presence of threedifferent classes closely related to the origin(station)of the samples.Thus it becomes feasible to proposefor each station a possible profile of the content in terms of the different parameters considered.