A Generalized Query Framework for Geospatial Reasoning

The ability to reason over geospatial entities using publicly available information is greatly enhanced by the abundance of data sources on the Internet. Traditional data sources such as satellite imagery, maps, gazetteers and vector data have long been used in geographic information systems (GIS). However, incorporating non-traditional sources such as phone books and property tax sites brings to light integration issues not previously considered. The ability to answer different types of queries about geospatial entities further complicates the problem. For example, assume we are provided with an image and consider the query “What is the name of this street?”. There exist many different ways of answering this query depending on the types of data sources and operations available to a system. Three possible plans which answer this query are shown in Figures 1, 2, and 3 . For each plan, we are provided with the latitude and longitude of the center of the image and a point identifying which street we are interested in. The first plan would query a vector data source1 which provides a street name(s). The second plan queries a gazetteer2 which returns a set of named points. These points could be used to query an online phone book data source for a street name(s). Finally, the third plan retrieves an image from an image data source, overlays this image on the map, and extracts the characters corresponding to a street name(s). These plans show there exist components that can: produce and pass new data to subsequent components in the plan, produce an incomplete set of data, or produce no data at all. Therefore, when constructing an execution plan, a system needs to consider not only the operations available to it but also the data produced during execution. There is no guarantee that a valid execution plan will produce a valid query answer. The notions of contingency planning or dynamic plan refinement need to be incorporated into the planning system.