Geodetski list, Vol. 66 (89) No. 4, 2012.
Stručni rad
Methods of Artificial Intelligence used for Transforming a System of Coordinates
Józef Gil
; Institute of Building Engineering, University of Zielona Góra, Zielona Góra, Poland
Maria Mrówczyńska
orcid.org/0000-0002-4762-3999
; Institute of Building Engineering, University of Zielona Góra, Zielona Góra, Poland
Sažetak
The article discusses the problem of transformation in the form of the function f:R2>R2 as dependence between the coordinates [x,y] of the original system and the coordinates [X,Y] of the secondary system. The task of the transformation of [X,Y]=f(x,y) technically understood as the transformation of a system of coordinates has been solved by means of the feed-forward neural networks of the sigmoidal type, radial neural networks, recurrent cascade neural networks, and neuro-fuzzy systems with the use of the Takagi-Sugeno-Kang model. The numerical procedures applied make it possible to obtain a level of accuracy of the task equivalent to the cartographic accuracy of pictures in the Spatial Information Systems.
Ključne riječi
feed-forward neural networks; fuzzy networks; coordinate transformation; geodesy
Hrčak ID:
102917
URI
Datum izdavanja:
21.12.2012.
Posjeta: 1.154 *