Vague Spatial Data Types

Naturally occurring phenomena in space often (if not always) cannot be precisely defined because of the intrinsic uncertainty of their features. The location of animal refuges might not be precisely known, and the path of rivers might be uncertain due to water volume fluctuations and changing land characteristics. The extension of lakes can also change and thus have uncertain areas. All these are examples of vague spatial objects. The animal refuge locations can be modeled as a vague point object where the precisely known locations are called the kernel point object and the assumed locations are denoted as the conjecture point object. The river paths can be modeled as vague line objects. Some segments or parts of the path, called kernel line objects, can be definitely identified since they are always part of the river. Other paths can only be assumed, and these are denoted as conjecture line objects. Knowledge about the extension of lakes can be modeled similarly with vague regions formed by kernel and conjecture parts. Figure 1 illustrates the examples above. Dark shaded areas, straight lines, and black points indicate kernel parts; areas with light gray interiors, dashed lines, and hollow points refer to conjecture parts.