LINE GENERALIZATION BY REPEATED ELIMINATION OF POINTS

This paper presents a new approach to line generalisation which uses t he concept of 'effective area' for progressive simplification of a lin e by point elimination. Two coast-lines are used to compare the perfor mance of this, with that of the widely used Douglas-Peucker, algorithm . The results from the area-based algorithm compare favourably with ma nual generalisation of the same lines. It is capable of achieving both imperceptible minimal simplifications and caricatural generalisations . By careful selection of cut-off values, it is possible to use the sa me algorithm for scale-dependent and scale-independent generalisations . More importantly, it offers scope for modelling cartographic lines a s consisting of features within features so that their geometric manip ulation may be modified by application- and/or user-defined rules and weights. The paper examines the merits and limitations of the algorith m and the opportunities it offers for further research and progress in the field of line generalisation.