To begin the inference process, the system needs input fuzzy sets. There are 2 cases :

• Raster inference: Input fuzzy sets are defined from values of the raster maps.
  Assume x the real value of the point being processed, the input fuzzy set is expressed by:
  

(x-3s, x-s, x+s, x+3s, 1-q)

          with s step of the universe and q ( ) quality of the map.

• Vector inference: Input fuzzy sets are the linguistic terms of the linguistic variables. If vector maps are built from raster map, geographical objects are built with linguistic term definitions.