3. Deconvolution & Sharpening

3.1 General

The Deconvolve & Sharpen tab contains all the necessary controls and sliders to control the sharpening settings. It is essential that you get these settings correct, please take the time to find the ones that are optimal for your gear setup. Besides the Telescope aperture, focal ratio, barlow lens and camera, also the seeing conditions may require a different value for the radius. 

3.1.1 Sharpening mode

The sharpening works in 2 different modes: RGB and Luminance mode.

3.1.1.1 RGB mode

RGB mode applies sharpening on the 3 channels Red, Green and Blue. This mode often results in the best color diversity, but it may introduce color noise.

3.1.1.2 Luminance mode

Luminance mode only applies sharpening to the Luminance channel. The Luminance channel  only works on the light intensity of the image while leaving the colors unchanged. The advantage of this is that sharpening will not introduce any color noise. The downside is less color richness in the image. This mode will work best on object surfaces that don't have a lot of color diversity such as the moon or the sun. 

LuckyStackWorker also provides additional checkboxes to compose the luminance channel. You can decide to drop a channel if it has more noise or a particular artifact is only visible in a certain channel. Finally you can also drop the whole color channel, effectively making the image an monochrome image.

3.1.2 Blend raw

With the blend raw slider you can mix the original unsharpened image stack with the sharpened result. This may give a softer effect, reduce the contrast which may have become too high from applying the sharpening and also reduces the noise. 

3.2 Deconvolve

There are 2 methods available to apply sharpening: deconvolve & sharpen. And you can also combine them. In that case first deconvolution is applied and then the sharpening. Deconvolve is based on the Wiener  Deconvolution and works with a so-called Point Spread Function (PSF). This is basically an image representing the dispersion of a singular point of light is dispersed by the atmosphere and the scope. In case of astrophotography applications  such a singular point is easily obtained by an image of a star. The PSF is used to reverse the dispersion effect by iteratively applying the algorithm. For more information see the following blog article that explains this in detail. Wiener Deconvolution is a very effective way to apply sharpening and it generally results in less noise and more contrast as compared applying traditional sharpening methods (Wavelet, Unsharp mask). But this can also depend on the conditions. 

3.2.1 PSF

When pressing the PSF button a new panel appears that facilitates the selection or creation of the PSF. There is a separate PSF kept inside the LSW storage for every object profile.

3.2.1.1 Custom PSF

In the ideal situation you should have made a star image (as described in the blog article referred above) which can then be opened and used. To do so press the  "Load custom PSF" button to select and store it. The deconvolution is then applied immediately based on that PSF and the chosen number of iterations. (see 3.2.2)

3.2.1.2 Synthetic PSF

In case you have no star image available that was made on about the  same moment when the images where shot it is still possible the generate an artificial (synthetic) PSF image. The following parameters need to be used in that case.

Airy disk radius

By adjusting this slider the size of the airy disk expressed in pixels on the sensor can be controlled.

Diffraction intensity

Adjusting this will change the visibility of the diffraction pattern. The pattern depends on the type of telescope used and the obstruction percentage of the mirror. An SCT will clearer diffraction rings visible than a Newton or a Refractor.

Seeing index

Use this slider to control the effect that the seeing has on the star image. The seeing index corresponds to the seeing index used on the well known Meteoblue seeing forecast. A lower value means worse seeing, which will have the effect of blurring the star. The highest value (5) means the seeing was optimal.

3.2.2 Deconv. iterations

This sliders allows adjustment of the amount of Wiener Deconvolution iterations applied. Note that the higher the chosen number the longer it will take to apply the algorithm.

3.3 Sharpen

The sharpen option can be enabled to apply the more traditional sharpening that is based on iteratively reversing the blurring effect. 

3.3.1 Radius

Radius can be set to lower values (below 1) for lower focal distances or cameras that have large pixel sizes (5 micron), and higher values for larger focal distances or when using a camera with small pixel size. The correct value for radius thus depends on the sampling ratio of the recordings. If the image was undersampled or the sampling was spot on, it is best to keep the radius low (1 or lower). And in case the recording is oversampled it means that the radius value needs to be higher in order for the sharpening to be more effective (and introduce less noise).

3.3.2 Amount

The amount slider defines the strength of the sharpening algorithm applied. Higher values will result in better details and higher contrast, but may also bring forwards the amount of noise. The noise can be suppressed by using the Denoising controls, described in section 2.3.

3.3.3 Iterations

The number of iterations can vary based on the quality of the recording, low values (1-2) for noisy image stacks, and higher values for low noise image recordings (usually bright objects such as the moon & sun). Multiple iterations can give sharper end results and finer details, but may also increase the noise.