ComputingComputer ScienceGeoinformaticsCartographyComputational GeometryComputer VisionCartographic GeneralizationGeometry ProcessingMap GeneralizationMappingShape AnalysisSpatial TransformationsAutomated GeneralizationAutomated Map GeneralizationContour AnalysisContour Line GeneralizationGeometric GeneralizationLine GeneralizationObjectsSkeletonization
Much work has been done in the last few years on various aspects of automated map generalization. However, many of these efforts have been directed to the generalization of individual objects- houses, roads, etc.- while ignoring the spatial relationships between the objects themselves. This problem is particularly apparent where map generalization includes the simplification of contour lines (which may overlap as a result, if no spatial relationships are preserved), or the question of preserving the relationships between different linear features (roads, rivers, railways) that must be cartographically displayed (even if displaced) in the smaller-scale maps after generalization. We here propose a new generalization method based on the medial axis transform, or skeleton. This new technique, as opposed to other traditional generalization methods, has the advantage of preserving the topological structure of the object, which eliminates any risk of overlapping of neighbouring objects. It is thus especially appropriate for the generalization of contour lines, among other things, where there may be high density and proximity of linear elements.
12
A transformation for extracting new descriptors of shape
H.E. Blum · Medical Entomology and Zoology · 1967 · 1.7K citations
Models for the perception of speech and visual form
Robert S. Brubaker · Journal of Communication Disorders · 1968 · 1.3K citations