期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2010
卷号：XXXVIII Part 4
出版社：Copernicus Publications
摘要：There is a constant effort by the data providers to maintain up-to-date geospatial databases. Multi-temporal coverage of the same geographic area using aerial and satellite images and airborne laser scanning data is the most common approach for directly generating and updating geo-databases from raw source data. Geo-databases can also be updated indirectly from multi-source multi-scale up-to-date geo-databases. The patterns of spatial changes in the existing data sets are detected and determined by determining mismatches in the state of a feature over different times. For the determination of spatial changes it is essential that the multi-temporal multi-source data sets are co-registered by mapping one data set over the other using appropriate transformation functions. Invariant corresponding features are matched between the two data sets for the determination of the transformation parameters. The methods and quality of data alignment are important in order to reduce registration errors as any difference is interpreted as potential change. In the actual determination of spatial changes factors such as level of change detection, types of data used for change detection, level of processing, types of changes, and turn-around time all influence the method applied. The change detection methods between existing and new features are based on approaches, which depend on the data types, level of pre-processing and on proximity and/or similarity measures. In this paper, using examples, we present methods and common practices applied for change determination in vector databases and identify trends for future approaches. Emphasis is on the use of automated methods due to the increasing data volumes and new types of data
