ERROR PROPAGATION THROUGH MAP OPERATIONS

In image processing and geographic information systems, a new map is c onstructed by carrying out a sequence of operations on a set of source maps. These operations typically include the adding, ratioing, and ov erlaying (or buffering) of two or more maps. But each source map may c ontain error. There may be error associated with measuring attribute v alues and in specifying the location (or position) of attribute values . This article investigates through theory and simulation the effects of different types of source-map error on both the size and pattern of errors in the resulting map and examines how these effects also depen d on the spatial structure of the true source maps.