Computer Science · Computer Vision and Pattern Recognition
Vario-scale data structures supporting smooth zoom and progressive transfer of 2D and 3D data
P.J.M. van Oosterom, Martijn Meijers
See on the mapAbstract
This paper introduces the concept of the smooth tGAP structure represented by a space-scale partition, which we term the space-scale cube.We take the view of 'map generalization is extrusion of data into an additional dimension'.For 2D objects the resulting vario-scale representation is a 3D structure, while for 3D objects the resulting vario-scale representation is a 4D structure.This paper provides insights in: 1. creating valid data for the cube and proof that this always possible for the implemented 2D tGAP (topological Generalized Area Partition) generalization operators (line simplification, merge, and split/collapse), 2. obtaining a valid 2D polygonal map representation at arbitrary scale from the cube, 3. using the vario-scale structure to provide smooth-zoom and progressive transfer between server and client, 4. exploring which other possibilities the cube brings for obtaining maps having non-homogenous scales over their domain (which we term mixed-scale maps), and 5. using the same principles also for higher dimensional data; illustrated with 3D input data represented in a 4D hypercube.