Combining Orthogonal Range Search and Line Simplification Using Priority Search Trees

Abstract

This thesis presents a method for combining orthogonal range queries and line simplification in 2D geographical data. Asking for data representing a specific area and have it displayed in an appropriate resolution is one of the basic operations of a geographic information system. In most applications today, the data is stored in several layers of different resolutions, and when a query is performed the approximation of the resulting area is created by querying in the correct resolution layer. The combined solution proposed here is based on the priority search tree (PST) developed by Edvard T. McCreight in the mid eighties. The PST is a data structure used for performing semiinfinite range queries. It was realized that its properties made it suitable for answering stabbing queries as well. This ability to answer stabbing queries is what has been utilized and developed further in the data structure presented here. To solve the line simplification part of the problem, an algorithm presented by Zhilin Li and Stan Openshaw in 1992 was used. This algorithm creates approximations by placing a grid over the map, and selecting the points where the original line intersects the grid lines. We realized that this problem could be solved as multiple stabbing queries. In the combined solution presented in this thesis, the PST is incorporated in another, two-level, data structure. A query in this new data structure returns an orthogonal range which is an approximation