Previous     Next

We now enter the realms of specialised image processing programs like Adobe Photoshop (PS) to bring the full depth of detail still hidden in the pre-processed image. There are, of course, other programs available, but I only know PS and will stick to that! There are down-sides to PS - it's expensive and complicated to learn. However, once you pick up a few tricks, it is amazingly powerful.

Before we started with PS, we need to download a free plugin to allow PS to handle FITS files. The plug in is called FITS liberator and can be downloaded here, as well as instructions on how to install it. Once the plugin is in place, this is the screen you will see when you ask PS to open a FITS file: 

FITS Liberator will guess at where to set the black point and white point (the two vertical sliders on the histogram). It will usually crop data, which I don't want at this stage, so I move the black point (the left slider) to 0 (or enter 0 in the Black Level box) and move the white point to the right, past any data points in the histogram. That way I can be sure I'm not prematurely discarding useful data.

What we are trying to achieve in the next steps is a broadening of the histogram. Currently the histogram is skewed to the left, which means all the data is compressed into a relatively small and dark 'window'. By using PS Levels and Curves we can gradually spread out the data in the histogram. This is called 'stretching'. Levels allow us to move the black and white points and perform a linear stretching. But you wil rapidly discover that is insufficient - we need the power of the Curves to allow us to selectively stretch and emphasise the interesting (typically faint) parts of the image without over-emphasising very bright or very dark features.

When the image is first opened, it looks washed out and there's little detail in the target, in this case the galaxy M77.

Invoke the levels function (Control-L) to perform some initial linear stretching (note the black line at the far left of the histogram).

We now have more detail showing. The black line at the far left of the histogram was due to a black border introduced during the pre-processing and stacking stage (due to the image shifting slighty during the 2 hours of exposures). This is cropped out of the image.

You may want to do another Levels correction to nudge the black point closer to the left of the histogram. 

Now we employ Curves (Control-M) to selectively stretch the interesting parts of the image (the fainter detail in the galaxy's arms, for example). This is a bit or trial-and-error at first, but you get a feel for what works best. Here, an 'S' shaped curve is stretching the fainter detail in the peak of the histogram whilst anchoring the brighter data (stars). You can perform the Curves several times; sometimes another Levels is required to darken the background my moving the black point to the right.

We are not getting a reasonable image, good detail in the galaxy with a fairly dark background. There will almost certainly be defects and there are numerous tools in PS to deal with them which are beyond the scope of this basic introduction. On my links page you'll find reference to Practical Astronomer and AstroPhoto Insight, two magazines which have run excellent series of image processing tutorials. Also Don Waid has produced a number of PS tutorials here.

These basic steps have shown how to release the hidden detail in an image. The image we've used was a monochrome Luminance image. You may also have the colour channels (R,G,B) and you will follow a similar set  of steps for each.

Once each of L, R, G and B channels have been processed, you need to combine them. The following is a short summary of the steps involved. If you are new to PS this may not make a lot of sense and I will in time expand on these instructions.

In PS, first step is to create a new blank RGB image. The easiest way to do this is pick one of the colour channels, press Control-A to select the entire channel image and then Control-C to copy the image to the clipboard. Then File->New Image will take the size of the blank image from the image in the clipboard. Make sure you select an image type of RGB and press OK. You will then have a blank RGB image.

You can select each channel from your actual image, Control-A, Control-C to copy first the R channel to the R channel of the blank image, then the B channel, etc.

Next we create a Luminance layer. On the newly created RGB image, add a new layer, with a layer type of Luminosity and set the Opacity to around 70-80%. Now, the human eye will take all the data from the Luminance layer and blend in the colour data, given the illusion of a detailed coloured image. In fact, I usually spend a lot of time sharpening up the detail in the luminance and actually Gaussian blur the RGB data. The eye still sees detailed colour imagery. Sometimes though, the colour data can get a bit washed out by the Luminance, so I usually boost the colour saturation. Without the Luminance overlay, the RGB data will look gaudy due to the saturation, but adding the Luminance tones it down to a more natural look.

Other tools & techniques

Finally I should mention two toolsets I use so much during my PS processing. Firstly, to deal with the gradients in my images (due to vignetting or the moon), I use Russ Croman's excellent GradientXTerminator. It's free to try, but you need to buy it after the trial period. Modest price ($49.95), very much worth it. Secondly there are a whole bundle of tools and tricks in Noel Carboni's Photoshop Actions. Again modestly priced ($19.95), but some fantastic widgets in there. Check them out.

Previous     Next


Make a Free Website with Yola.