Preparing an LRGB image
The most frequent approach to create good looking colour images of astronomical targets is the LRGB technique. Therein, a set of RGB images get combined with an unfiltered
Luminance image. The RGB images contribute only the colour information, whereas the brightness information comes from the unfiltered image. If that combination is done with the wrong technique, then the resulting images appear rather flat in colour. Below I describe a method that conserves the colour of the RGB image when merged with the L channel.
Combining the RGB image
The individual R, G and B 16-bit TIFFs are combined into an RGB image with
Channels -> Merge Channels -> Mode: RGB colour
from the Channels window menu. To this end, the three files must be opened already, otherwise the Merge channels option is greyed out.
In the next step, the curves are steepened until a background level of about 20 is reached in RGB. This can involve partial masking of brighter regions to avoid saturation.
Colour noise reduction and colour saturation
To avoid colour noise, do Filter -> Noise -> Despeckle over the RGB image.
If the RGB appears to have low colour saturation, the latter can be increased by 20% to 40% (Image -> Adjustments -> Hue/Saturation). In case of bright emission nebula the colour saturation is usually already high enough.
Combining the RGB with the Luminance channel, using the Lab format
Copying the contrast enhanced Luminance channel over the RGB image with Blending mode: Luminosity results in distorted or unpleasant colours. A way that works much better is to
- convert the RGB image into Lab format (Image -> Mode -> Lab colour)
- open the grey-scale Luminance image
- split the Lab image into its individual channels, through the Channels window
- from within the Channels window, do Merge channels -> Mode: Lab colour and then select the Luminance image as the Lightness channel. It can happen that the resulting image looks a bit darker than expected, in which case the curves of the lightness channel can be steepened a bit further.
- Lastly, convert the image back into RGB
