Visualisation with 'intelligent' assistance: a methodological approach

A specific approach is described to the elaboration of a new-generation three-dimensional image-synthesis system built around a language taken as the main communication medium between the different modules of the system. The outlines of the subject are deliberately limited to the data-acquisition and object-construction aspects. Theoretical and methodological deductions are drawn from the possibility of converting different data used in the various modules. These are illustrated with examples of visualisation of urban scenes. A common language medium is proposed to support the 'intelligent' behaviour of the system, which starts to work once a critical threshold is reached: first, through a learning module based on a user -- system dialogue; then through the further development of the system in a progressive self-learning mode.