Simplifying Cartographic Boundaries by Using a Normalized Measure of Ambiguity

| This paper presents a method for smoothing of noisy carto-graphic boundaries. The resulting smoothed description preserves the basic shape's structures of the contour while smoothing out noise as well as un-wanted detail. The performance of any simpliication method relies on the quality of the levels of smoothing that represent the structures present. In our formulation the most signiicant levels that best describe a given contour are determined by examining a measure of redundancy between each two successive smoothed curvature graphs along the boundary. These levels correspond to the locations of the local minima of such a measure of redundancy. Experimental results are presented showing that complicated planar curves like cartographic boundaries could be smoothed in high accuracy. It is also proven that an exaggerate noise added to the boundary can be handled by our method. The simpliication is therefore reliable, robust, as well as very eecient.