SkyDude
Astronomy
PROCESSING
Software Selection
For a long time I
used Nebulosity for pre-processing (calibration, stacking, etc) and
PhotoShop CS2 for post-processing. But due to their limitations it was
time to upgrade. Several packages were exhaustively tested, including
the latest version of PhotoShop, StarTools, Astro
Pixel
Processor, and PixInsight which became the winner.
PixInsight Pros:
- Includes
pre-processing as well as post-processing
- Seems
to have solutions for just about any astro-imaging problem
- Commonly
used - it's almost a standard
- Not
as difficult to learn as you may have heard
PixInsight
Cons:
- Computer
resource hog - check system requirements before using
- Off
beat user interface
I
still use Nebulosity for image capture, and PHD2 Guiding for guiding.
Pre-processing
Workflow
General
Notes:
- Instead
of using the Weighted Batch Processing script that many people
use, I prefer to perform each pre-processing step separately.
- Occasionally
"Blink" is used to delete sub-frames that show obvious defects such
as airplane trails or guiding bumps.
- Master-dark
and master-flat frames were prepared earlier - I don't recreate them
after each imaging session.
- Removing
dark borders at this stage allows post-processing auto-stretch
processes such as MaskedStretch
to work
better.
Dialog
entries are
default unless otherwise stated.
ImageCalibration:
- Performed
on each L,R,G,B group separately
- Master
Dark = yes, Optimize = no
- Master
Bias = no
- Master
Flat = yes
CosmeticCorrection:
- Performed
on all
sub-frames simultaneously
- Use
Master Dark = no
- Use
Auto Detect = yes
- Hot
Sigma = on
- Cold
Sigma = on
StarAlignment:
- Performed
on all
sub-frames simultaneously
- Generate
Drizzle Data = off
ImageIntegration:
- Performed
on each L,R,G,B group separately
- Image
Integration / Combination = Average
- Pixel
Rejection (1):
- Rejection
Algorithm = Winsorized Sigma Clipping
- Generate
Rejection Maps = off
Crop:
- Margins/Anchors
- set
as required to remove dark borders
- Performed
on each L,R,G,B color-channel image, using same Margins/Anchors
Post-processing
Workflow
General
Notes:
- MaskedStretch
is used because it seems to keep stars from becoming overpowering,
compared to other intensity-stretching processes.
- Performing
an intensity-stretch before LRGB Combine results in less color
artifacts.
- To
avoid Deconvolution artifacts, color
noise is reduced first, and a lightness_mask is used.
- Occasionally
CloneStamp is used on the object_mask to remove artifacts or
missed-stars caused by StarNet2.
Dialog
entries are
default unless otherwise stated.
MaskedStretch:
- Performed
on each
L,R,G,B color-channel image
- All
dialog values = default
LRGB
Combine:
- All
dialog values = default
ACDNR
(Adaptive Contrast-Driven Noise Reduction):
- Purpose
- to reduce
color noise
- Lightness:
- Chrominance:
- Apply
= on
- Lightness
Mask = on
Copy
CIE-L Channel - to make lightness_mask
Deconvolution:
- mask
= lightness_mask
- All
dialog values = default
CurvesTranformation:
- Purpose
- reduce background glow, increase overall intensity
- Mask
= none
- Dialog
values = adjust as required
StarNet2:
- Purpose
- optional step to make mask of object (galaxy, nebula) with
no stars
- Copy
lightness_mask, rename the copy to object_mask
- Perform
StarNet2 on object_mask
- All
dialog values = default
CurvesTranformation:
- Purpose
- enhance object/s of interest only
- Mask
= lightness_mask or object_mask
- Dialog
values = adjust as required
ACDNR
(Adaptive Contrast-Driven Noise Reduction):
- Purpose
- to reduce
overall noise
- Mask
= none (except for the Lightness Mask built into ACDNR)
- Lightness:
- Apply
= on
- Lightness
Mask = on
- Chrominance:
- Apply
= on
- Lightness
Mask = on
MMT
(MultiscaleMedianTranformation):
- Purpose
- to enhance small detail
- Mask
= none
- Mode
= MWT
- Layer1
= 0.3
- Layer2
= 0.3
Special
Processing Techniques
. . . coming soon . . .
Comments
welcome: mark.eby@twc.com