The context related to each production rule is computed for each point using fuzzy sets theory, fuzzy logic and mathematical morphology.

For example, the premise X "near roads" implies

• road detection in the image or road map access in the G.I.S.,

• real distance computation (using Digital Elevation Model) and finally

• attribution of a likelihood degree .

The conclusion of a rule can be represented by a combination of intersections, unions and negation of premises. In this case, the following computations are done previously to produce the realization degree for combined premises :