Large vario-scale datasets

Abstract

In Chapter 3 the focus was on vario-scale data structure description. This was extended in Chapter 4, where generating better content for this structure was investigated. It showed how the structure has been developed and used in practice, and current technical limitations. One of them is processing really massive dataset with records in order of millions which do not fit in the main memory of computer. It is a notorious and challenging problem. This is especially true in the case of map generalization, where the relationships between (adjacent) features in the map must be considered. Therefore, this chapter presents our solution for automated generalization in vario-scale structure based on the idea of subdividing the workload according to a multi-level structure of the space, allowing parallel processing. More specifically: Section 5.1 specifies our goal. Section 5.2 presents related work and other options to handle large datasets. Section 5.3 explains the principles of our method in more detail. In Section 5.4 modifications of the process specific for road network generalization are introduced. Statistics and a test of real dataset with more than 800 thousand objects are given in Section 5.5, followed by conclusions and the future work related to processing large datasets in Section 5.6.