1 de julio de 2000 Vol. 1 No.1 |
Abstract. We present an integrated technique for the completely automated digitization of raster maps. This technique considers in detail all three steps involved in digitizing raster maps; pre-processing, processing and post-processing. Current programs for vectoring maps using automated or semi-automated systems essentially carry out only one kind of operation; they follow discrete points along a curve. In our proposal, the automation problem is treated from a unified point of view, leading to the creation of software that is capable of recognizing a maximum number of cartographic images in raster maps. We propose some strategies for solving the problem, and illustrate with an example of an automated system to vector cartographic point objects. Key words. Geographical Information System, map raster processing, cartographic pattern, correlation, and neural network.
|
[ Este
número
]
|
Dirección General de Servicios de Cómputo Académico-UNAM |