The following classification is reached (table 28) with the degree of consensus h computed locally to each pixel.

The adaptive fusion of the four spectral bands MSS4 to MSS7 gives a result almost identical to that obtained by conjunctive fusion (53.68% of pixels correctly classified for adaptive fusion against 53.35% for conjunctive fusion). On the other hand, all the pixels were classified. This is due to the ``disjunctive fusion'' part of the rule which allowed to assign to a class the pixels which had not been classified by conjunctive fusion.

Adaptive fusion does not allow to obtain better results than simple conjunctive fusion on this example (except for unclassified pixels) because the spectral bands have a too strong degree of coherence. There are seldom conflicts to solve, so the ``conjunctive fusion'' part of the adaptive rule is used in 98.79% of the cases. So results are similar to those obtained by simple conjunctive fusion.

Table 28: Adaptive fusion of spectral bands MSS4 to MSS7.

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	151	459	32.90%
2	230	459	50.11%
3	122	306	39.87%
4	204	391	52.17%
5	268	459	58.39%
6	294	459	64.05%
7	333	459	72.55%
8	299	459	65.14%
9	198	459	43.14%
TOTAL	2099	3910	53.68%