1. Extract Lightness image from original using 'ChannelExtraction'.
2. Create a preview which includes various star sizes
GUIDELINES
- If noise or other non-stellar structures are included, raise the 'Noise threshold'.
- If large stars are not included, raise the 'Scale' amount.
- If needed, raise the 'Shadows' slider to further restrict selection to bright stars.
- Lowering 'Truncation' will also provide a stronger result.
- Deselect galactic cores using the Clonestamp process (Ch 17) or by adjusting the 'Midtones' slider further.
- Structure Growth
> Lower 'Large-scale' to make large stars smaller, increase for the opposite effect.
> Lower 'Small-scale' to make small stars smaller, increase for the opposite effect.
- Midtones
> Raise to exclude dim stars
> Lower to include dim stars1. Pick stars surrounding the areas of interest until around 120 entries appear in the table.
2. Sort list with increasing levels of MAD.
3. Generate a PSF model using the best ~50 stars using 'Export synthetic PSF'.
- Try to find around 50 stars (for a mono image) grouped as close together, at the lowest values of MAD, as possible.
4. Select all rejected stars and use the 'Remove' button at the bottom.
5. Select all the remaining stars, and click the Camera icon to generate a PSF.
M8_Ha
1. Clone the original image.
2. Rename is 'deconv_mask'
3. Apply a HistogramTransformation stretch based on the STF Auto-stretch.
1. Clone deconv_mask and rename it 'nr_mask'.
2. Apply this HDRMultiscaleTransform process to the deconv_mask
3. Clone the deconv_mask and rename it 'tmp'0.0001,0.003162,0.01,0.003162,0.0001,0.003162,0.1,0.316228,0.1,0.003162,0.01,0.316228,1,0.316228,0.01,0.003162,0.1,0.316228,0.1,0.003162,0.0001,0.003162,0.01,0.003162,0.00010.01,0.316228,1,0.316228,0.010.01,0.316228,1,0.316228,0.01Gaussian (5)1. Apply this process to the tmp image.
This step increases the contrast of an image for several different structure sizes. I like to use the LocalHistogramEqualization tool for larger scales and the MultiscaleMedianTransform for smaller scales. Apply this process to the tmp image.
1. Apply this PixelMath process to the deconv_mask image.abs($T-tmp)*$T + 0.1*nr_mask
1. Select nr_mask as the support image
2. Apply this process to the deconv_mask image.nr_mask1. Start testing settings using 10 Iterations.
2. The two main things to adjust are:
- Number of iterations (Normally between 20-50)
- Global Dark Setting (Normally between 0.01 - 0.05)
3. For images with no bright stars over bright areas you can disable local deringing.
4. You should be appling this process to the main image.star_maskPSF
1. If you find that the deconvolution isn't strong enough apply this process to the deconv_mask image before using it.I wasn't get very much deconvolution effect so I stretched the deconv_mask to make it slightly more permissive. Apply to the deconv_mask image.