Re: Colour image deconvolution

Signal processing in color images depends heavily on the type of images
and applications you are working with. Usually, in GIS applications,
image channels are split and processed separately as if you have 3 (or
more) single-channel images. On the other hand, color medical or
natural images may require processing in the "signal power" scope,
which means that they should be treated as one single-channel image. In
the second case, you can use a proper channel mixing model (weighted
average), process the intermediate greyscale image, and then apply the
reverse mixing model to "map" the resulting single-channel image into 3
(or more) channels.