Table 29: Adaptive fusion of spectral bands MSS4 to MSS7 and of out-image data: distance to roads, plateaux, valleys, slope orientation, slope steepness, crests, irrigation, DEM and distance to urban areas

Rates of classification obtained
Class	Number of pixels   correctly classified (A)	Number of pixels in samples      (B)	Rate of pixels     correctly classified (A/B)
1	342	459	74.51%
2	251	459	54.68%
3	84	306	27.45%
4	84	391	21.48%
5	146	459	31.81%
6	185	459	40.31%
7	62	459	13.51%
8	37	459	8.06%
9	237	459	51.63%
TOTAL	1428	3910	36.52%

The results are quite different from those of simple conjunctive fusion. The mean rate of recognition was 36.52% (against 22.89% for conjunctive fusion). In fact mainly classes 1 and 2 benefit from a significant improvement of their rate of recognition. That is the effect of disjunctive fusion. As we noticed in the previous chapter, disjunctive fusion often attributes after fusion the same degrees of membership for a pixel to several classes. In such an approach, our implementation assigns the pixel concerned, to the first of the candidate classes.

Let us notice that the problem of pixels unclassified with conjunctive fusion disappeared. Moreover, the rates of recognition of each class are better than in the case of simple disjunctive fusion. Thus, the conjunctive part of the adaptive rule worked when it was possible.

However, the image obtained is very unsatisfactory because the adaptive rule, in the presence of conflict, often used disjunctive fusion. In fact, if there is conflict between the sources and that the unanimity cannot be reached to assign a pixel to a class (unclassified pixels for simple conjunctive fusion), the disjunctive part of the adaptive rule is used. That occurs in 93.03% of the cases for the whole of the image during the fusion of the spectral bands and the out-image data.