Out-image information plays an important role in the forecasting, and therefore must be considered to improve the results.

These data may be represented by a map where each pixel is associated with coefficients calculated according to expert knowledge by means of rules. These coefficients are merged to the coefficients obtained from the previous stage in order to precise the consensus zones.

Several information-fusion methods are available for different kinds of problems. In our image processing application, we have used a weighted average to merge the coefficients resulting from different sources of information and thus, the influence of each source of information is dependent upon the provided weights. Finally, we obtain an image of coefficients of evolution, belonging to [0,1].