Image Processing Reference
bottom corner seem to lack in contrast. This area appears well-contrasted and thus,
is quite clear in the results of the other two solutions as shown in Fig. 10.2 b-d. These
techniques explicitly consider the contrast measure either from the input or from
the output perspective, which leads to such improvements over the other techniques.
Figure 10.2 c shows the result of the variational fusion technique. Although it is similar
to the result of the bilateral filtering-based technique, it lacks in sharpness, and thus,
smaller objects in the right side of the scene are not very clearly observable.
Higher values of variance
2 , and average gradient
g for the Bayesian and the
optimization-based techniques seen in Table 10.3 are in agreement with the corre-
sponding results in Fig. 10.2 . However, we observe similar values of entropy for
all the results, which indicate a similar amount of average information content for
all images. The fusion factors (FF) for all the presented techniques indicate their
usefulness towards producing results with higher contents from the input data.
Fig. 10.2 Results of visualization of the moffett 3 data from the AVIRIS using a the bilateral
filtering-based technique, b the Bayesian technique ( c
2013 Elsevier, Ref. ), c the variational
technique, d the optimization-based technique, e the three band selection technique, f the piecewise
linear function technique, and g the color matching function technique