I don't understand... Do you have six monochrome images and you want to combine them to one color image such that each of the six is still recognizable?

yes i have six monochrome images and i want to combine them in RGB images, the output will be with 3 channels, red, green, blue.

Well, if you convert one image to red and one to blue, then most people can grasp that any shade of purple means some overlap. Adding a third (e.g. green) complicates matters in the sense that the three individual images will not be recognized.

So, if the images overlap, it cannot be done.

If not a human but a computer should be able to separate them, then you can mix three images (four if you abuse the alpha channel). You could put six if you reduce the bit-depth of your monochrome images to 4 bits.

ImageJ (http://rsb.info.nih.gov/ij/ ) can easily merge up to 7 image channels into a single RGB image. You have control over the ordering and the color of the channels. It's also possible to use the OME (Open Microscopy environment) to produce similar results. ImageJ would be simpler to setup and use.

Thank you Mr. Lambert Zijp, yes i am agree with you that there is some color systems can not be recognized by human, but i want to make the best available solution for these color channels. By taking into account that each channel is indicate one color spectrum, and some color spectrums are not tied to some other spectrums, so we can neglect one or more of these spectrums in making the new RGB image channels. Is these true?

With reference to overlap and shading it depends to each channel separately, we can not control or find out the imaging conditions, the information not available for us, but we can do some pre-process per channel before mixing them.

Thank you Mr. H. Morris i mix the 6 channels with ImageJ but i do not know exactly how the channels be mixed, i need to figure out the formulation for this process, if you suggest some papers or researches, i will be thankful to you

Quote Osama: "By taking into account that each channel is indicate one color spectrum, and some color spectrums are not tied to some other spectrums, so we can neglect one or more of these spectrums in making the new RGB image channels. Is these true?"

Pure Red, pure Green and Pure Blue are not tied. Any other color is a combination of one or more of these channels.

If your monochrome images are 8 bits, then you can only store three of those images in a 24 bits RGB image without losing information. As said, by reducing your images to 4 bits, you can store six.

So: without loss it cannot be done.

Sure, you can semi-transparently overlay a color wash on top of a black&white image. Below is a radiotherapy example with a black&white CT slice and a BlackBlueGreenYellowRed dose distribution.

My point is that neither the original CT nor the original dose can be recovered from the combined one.

Yes Mr. Lambert Zijp i understand what you said about recovering the original images from the combined one, i am 100% agree with you. I do not said that i want to recover them from original, i just want to make a colored image from spectral channels to apply some processes on it, and the making process Involve the greatest amount of information as much as possible. thank you a lot for your attention sir.

Mr. Volker Jaedicke thank you for these references, its interesting to go deep in image colors and spectrum analysis. the spectral imagery is a wide area and your paper will be a good launch for me.