Image Processing Reference
all these bands are clubbed into a single color channel. Each of the groups have been
fused separately, and the three fused images are assigned to R, G, and B channels to
form a composite color output.
This chapter develops a fusion methodology that transforms the problem into an
optimization framework where visualization is the primary aim. Therefore, it focuses
on computing the fusion weights optimally in order to provide better visualization
results. The novelty of the discussed solution lies in the fact that the fusion process
is driven by the desired properties of the output image as opposed to the traditional
methods which are driven by the input. We have explained a multi-objective cost
function based on some of these image properties, and then discussed the solution
using the Euler-Lagrange equation. Being completely output-driven, this solution
does not require any assumption on how the weights should be related to the input
Our discussion also includes all typical constraints of pixel-based image fusion.
We have, thus, demonstrated how fusion can be represented as an unconstrained
optimization problem by introducing an auxiliary variable which leads to a compu-
tationally simpler and easier solution.