A general framework for spatial data inspection and assessment
Yiliang Wan,Wenzhong Shi,Lipeng Gao,Pengfei Chen,Yong Hua
DOI: https://doi.org/10.1007/s12145-014-0196-9
2015-01-01
Earth Science Informatics
Abstract:The quality aspects of spatial data are very important in the decision-making process. However, the quality inspection of spatial data is still dependent on manual checking, and there is an urgent need to develop an automatic or semi-automatic generic system for spatial data quality inspection. In this paper, we present a general framework that automatically copes with spatial data quality inspection based on various spatial data quality standards and specifications. The framework involves all descriptions of given spatial data, a data quality model characterized by quality elements, scheme batch checking and spatial data quality assessment based on quality control and assessment procedures. It is implemented in Unified Modeling Language with four main sets of classes: data dictionary, quality model, scheme checking and quality assessment. Accordingly, we have designed four structured Extensible Markup Language files for the framework to organize and describe the data dictionary, quality model, scheme check and quality assessment. It is very easy for users to describe the data requirements using the data dictionary file, and to extend the quality elements or check rules using the quality model file. Users can design the specified checks and quality assessment schemes without coding by configuring the scheme check files and quality assessment scheme files. The framework also incorporates a checking tool capable of solving the difficulties inherent in the diversity of spatial data quality standards and specifications. The proposed framework and its implementation, as a quality inspection system, will facilitate automatic multiple spatial data quality inspection and acceptance. As a result, the quality of diversified spatial data can be ensured and improved, which is extremely important in the era of spatial big data.