【作者】 曾衍伟；

【导师】 龚健雅；

【作者基本信息】 武汉大学 ， 摄影测量与遥感， 2004， 博士

【摘要】 随着信息技术的发展,地理信息系统(GIS)的应用日益广泛。在我国GIS发展初期,李德仁院士(1991)指出:如何建立GIS的质量模型、用什么尺度来度量GIS中的精确数据和非确定数据、定量数据与定性数据等是GIS发展必须解决的问题。GIS数据不确定性以及由此产生的GIS数据质量控制技术成了国内外GIS研究的热点问题之一。 近年来,空间数据位置不确定性成为研究热点,并已取得一些成果:点位不确定性采用一系列误差区间来度量,线位不确定性采用误差带来度量,采用构成面元边界线段的不确定性描述面元不确定性。以往研究的重点集中在建立位置不确定性模型及度量指标、遥感分类数据的不确定性等方面,对于如何在数据获取过程中进行质量控制以确保成果数据符合质量要求探讨较少,是值得深入研究的课题。 本文首先分析空间数据质量问题,归纳可能出现的质量问题,并以地形图扫描数字化为例,分析空间数据的误差来源;然后,建立统一的空间数据成果质量模型;接着,从GIS数据采集过程着手,研究如何在数据获取过程进行质量控制从而确保成果数据质量,以及研究空间数据质量指标的检验与评价方法,尤其是采用程序进行检验的方法:最后,开发相应的软件,实现空间数据质量控制与评价,从而建立完整的空间数据质量控制与评价技术体系。本文的主要研究成果包括: 1.分析了空间数据质量 分析了空间数据质量问题,并将其分为误差和错误两大类。针对GIS数据获取方式多、工序复杂,以地形图扫描数字化为例,分析空间数据误差来源,探讨各作业工序可能引入的误差、误差表现形式及其对成果数据质量的影响;通过误差分布检验,分析了地形图扫描数字化数据误差分布,为地形图扫描数字化数据质量控制提供依据。 2.建立了统一的空间数据质量模型 在提出的各种空间数据质量元素基础上,针对GIS数据格式、数据类型多种多样,目前尚无统一的数据质量模型,通过归纳、指标细分等方法,提出采用“精度”、“一致性”、“完整性与正确性”作为描述空间数据质量的一级质量元素,并详细列出了各个一级质量元素相应的二级质量元素、各个二级质量元素相应的三级质量元素,建立基于GIS数据内容的、统一的空间数据质量模型,作为GIS数据质量控制、质量评价的依据。然后,以统一的空间数据质量模型为基础,建立了数字线划图(DLG)、数字正射影像(DOM)、数字高程模型(DEM)的质量评价模型,作为产品质量检验与评价的标准。 3.提出了空间数据质量控制方法 提出了空间数据质量总体控制方法,并以地形图扫描数字化为例,探讨了如何从选择数据源、地图扫描、扫描图像几何纠正、图纸定向、屏幕数字化、数据编辑、成果数据位置精度限差规定、位置精度检测时检测点个数的确定等方面进行工序质量控制,确保获得高质量的DLG成果数据、应用相关软件生成高质量的DEM成果数据。 4.提出了空间数据质量检验与评价方法 首先,根据GIS数据质量模型,研究了需要检验的质量指标,探讨了不同质量指标的检查方法;进一步区分出可以采用程序检查的质量指标;最后,研究质量检查、质量评价的实现方法。主要成果包括:更多还原

【Abstract】 With the development of the information technique, the use of Geographical Information System(GIS) becomes broader and broader. At the GIS development beginning stage in China, Prof. Deren Li (1991 ) pointed that the following problems must be solved: how to build quality model of GIS, which index is used to measure accurate data and uncertainty data, quantitative data and qualitative data in GIS.GIS data uncertainty and GIS data quality control technique becomes a research issue.Recent years, positional accuracy of the spatial data is a research issue, and some research result is gained. Point uncertainty is measured by error intervals, line positional uncertainty is measured by error belt, area positional uncertainty is described by uncertainty of lines which form the boundary of the area. The former research emphases on the foundation of positional uncertainty and measurement index, remote sensing classification data uncertainty etc. Fewer discussion is made on how to control data quality in the process of data collection to insure data production quality to meet the quality requirements, which deserves deep research.At first, spatial data quality problem is analyzed and possible quality problem is summed up. Map scanning digitization is used an example to discuss the error source of spatial data. After that, an uniform spatial data quality model is built. Then starting from the GIS data collection process, how to control data quality in the process of data collection to insure data product quality is researched. Next, spatial data quality control and evaluation method is discussed, and the program check arithmetic is emphasized. Finally, the correspond soft of the spatial data quality control and evaluation is developed to control and evaluate spatial data quality. So a complete spatial data quality control and evaluation technique system is founded. The main research result is shown as the following:l.Spatila data quality is analyzed.Spatial data quality problem is analyzed, and is divided into error and mistake. Aiming at GIS data can be collected by much methods and working procedure is complicated, map scanning digitization is used as an example to analyze the error source of spatial data. The error source of map scanning digitization is detailedly analyzed. Possibly imported error, error representation form and its impact on the accuracy of data product is probed. Error distribution of map scanning digitization is analyzed by the way of error distribution test to give the basis of map scanning digitization data quality control.2.An uniform spatial data quality model is founded.Aiming at much GIS data format and data type, and there is not an uniform spatial data quality model, on the basis of GIS data content, quality elements Accuracy, Consistency and Completeness & correctness are suggested to measure spatial data quality by the way of induction and subdivision. And data quality elements and their quality subelements is given in detail. As a result, an uniform spatial data quality model is advanced, which is used as the basis of GIS data quality control and evaluation. Using this data quality model, data quality evaluation models of DLG,DOM and DEM are founded, which are used as a standard of product quality testing and evaluation.3.Spatial data quality control method is advanced.Using map scanning digitization as an example, general spatial data quality control method is advanced. To insure the data product to meet quality requirements, the following method were discussed as process quality control: select data source, map scan, geometrical rectify, map orientation, screen digitization, data edit, positional accuracy limitation defining of data product, testing point number of positional accuracy test, etc. By this way, DLG data with high quality can be gotten, and DEM data with high quality can also be gotten using correlative DEM making software.4.Spatial data quality control and evaluation method is suggested.At first, data quality test index is researched according to GIS data quality model. And test method of the different quality index is discussed. Next, quality index which can be tested by program is distinguished. Finally, quality test and quality evaluation method is researched. The main research results is introduced here.? Using template match technique to check attribute data and metadata.? Using special topological relation test to check the topological consistency.? As for data cartography rule, element relative relation among different data layer, edgematch of different data file, etc., rule base is founded to check data consistency of different data by the way of rule defining by user.? Using mathematical method to check positional accuracy.? Vision model of man eye is established using experiences, the statistics value of the testing image is computed. They are compared to check image quality.? To check completeness and correctness automatically using the specific demands of the specification and technique design.? To represent error information in different ways, especially by means of graphic to become convenient for error affirming and editing.? To evaluate data quality automatically with the comparison of quality standard database and data quality testing result database.5.Spatial data quality control and evaluation software is developed.New spatial data quality test and evaluation idea is advanced. The complicated spatial data quality characteristics is abstracted as rule and model, which is used as the theoretic basis to check and evaluate the different data quality by means of program.Based on above research results, two universal software is developed, which is named as Vector Data Quality Control and Evaluation System(DLG-Checker) and Image Data Quality Check and Evaluation system(DOM-Checker) respectively. They have the following merits:? The graphic and image operation environment are developed by advanced language and database, which has independent knowledge property right.? They can be used to control procedure data quality and data product quality , check the quality data which will be imported database.? Spatial data quality control rule base is founded according to the knowledge, such as quality control experience, data quality standard cartography rule, etc. On the basis of rule and model, the test scheme is made flexibly, so test content and evaluation standard for different scale and different quality demand can be adjusted conveniently to meet the requirements of different project data quality control and evaluation. Besides this, these soft can check data quality in different type of format. In short, they can be used universally.? Using database technique to manage and update spatial data standard, the result of data quality check and evaluation.? Provide data edit function to adjust the error during data quality control process.? Facing digital surveying and mapping production, these soft can be used to check and evaluate the quality of different data by means of adjusting test scheme.Single function test and comparative test of these soft shows that these soft can be used universally, testing and evaluating result is reliable, quality control productivity is improved greatly.更多还原