Notebook

2022

120 pages to get started with Pixinsight

Simple treatments with photos with/without DOF

Nicolas Catrix

h ps://[Link]/carnetderouteastro
[Link]

With the support of

Hello!

If you are reading these few lines, it means you are looking for a support that allows you to
discover the main features of Pixinsight. A fan of Photoshop, I had a bit
I'm having a hard time getting used to Pixinsight. The latter is a true tool dedicated exclusively to
astrophotography. At first glance, the interface is confusing for the novice but with
a little experience, it brings a real gain in the quality of our images. This pre-
The presentation is not intended to be a reference but simply a possibility of...
a little processing among many others... This tool aims to be a guide to help you
save time in the approach to Pixinsight.

The set of photos are not of good quality to show you the advantages.
undisputed of this software. Imagine the result with stunning pictures with autoguiding...

Happy reading!
Nicolas

NGC 6995 P04

NGC 281 P42
M66 P52
NGC 7023 P75
NGC 7023 P99
NGC 6979 P107
 SUBJECT: NGC 6995
CAMERA: ATIK HORIZON 2
20 x 90s

FILTER: IDAS NB1 (tri-band)

WITHOUT AUTOPILOT
WITHOUT DOF
 THE PIXINSIGHT INTERFACE
Pixinsight is a tool that can...
saint and dedicated to treatments
astrophotographs. Here I
will present this or
the processing of a photo
colors (CMOS).

Pixinsight is not 'my-

gique" because it will not correct
the defects of tracking, of packages
My image Process acf
macron or even times
too short poses... but
he will undoubtedly improve
your photo. Let's discover
together the interface of this
software.

In my image, I can it
view in large or in pe-
or still in full size
to observe in detail...

Maximum size on my screen

Reduce the size on my screen
Actual size of my image

For the active process, its pro-

properties are presented under
the shape of a window with
tabs.

There are 2 ways to apply

this process has an image:
Slide the triangle on a
image
Click on the square to
apply on the active image.

Reset to default values

 THE PIXINSIGHT INTERFACE
Processes can be
pinned to the right what
allows to find them more
easily and also to deter-
to mine a specific order for the
treatment...

To open the window of

process, I click on:

To streamline the process, I

click on:

To open an image, we
click on :

To manage the masks

(hide, see, select or
still delete) :

To create a preview
on one side of the image
(corresponds to the green square) :

Menu where the in-

set of useful processes
pre-processing, processing
cosmetic...
 CASE STUDY: COLOR PHOTO WITH 3-BAND FILTER
To obtain a final photo
very correct, it is important
to select the best
their photos from the shooting.
To do this, I will carry out a
first visual selection
with the or OUT PROCESS /Blink

This or that is a kind of

photo viewer... For
select my photos, I
click on :

Once the photos are integrated,

It is enough to use the arrows
to verify manually
if a photo contains
traces of satellites or a
cloudy passage

For a large number of

photos, I can make a
automatic reading all
half-seconds and throw
the soul

If I find a photo that

I do not wish to keep,
I can move it into a
another directory (for example
"trash") by clicking on:
 SELECT THE BEST PHOTOS
After a first selection
in a visual, it is possible to
select the photos of fa-
more rational thanks to the
ProcessSubframeSelector

Images need to be added to

study by clicking on:

First, to measure our

photos, we choose:
Rou ne: Measure Subframes

Then, it is necessary to indicate

some information
notably in 'System
Parameters at the level of
"Subframe scale" which we
can calculate here: h ps://
[Link]/echanllon-
swimming-and-visual-field/

It is enough to enter the focal length of

the instrument (mm) and the size
of a pixel (µm).

Then, we can launch the

process by clicking on:
After a few seconds
of analyses (according to the number
of photos), we get a
new window with a
graph...

In the table, I have the list

of my photos. In the
graph below, I have the
measurement of the FWHM of my
photos. The lower it is,
the more the stars will be born
and small.

This is where I select

the photo with the FWHM there
lower because it will serve me
reference photo for
the alignment... (Here it is the
photo 1).

Now, I am going to delete-

Put the photos where the FWHM
is bad and therefore where she
is the highest.
For this, at the level of the gra-
you just have to click on
the points (corresponding
in the photos) that I do not sub-
don't keep in touch
One will notice that
the points are transformed into
cross...

Finally, I will be able to save.

keep the 'good' photos.
I am switching to Rou ne mode:
Output Subframes.

In the 'Output Files' area,

I indicate the directory where
my "good" photos will be
mises. Ensuite, je valide en
clicking on :
 DEBAYERIZE THE PHOTOS
I will open my photo of
reference on Pixinsight. She
appears very black and it is
normal.

So that my photo is
visible, I am going to use the
Process /ScreenTransfer-
Function

This or that simply allows

to visualize the photo without
modify the pixels.

By clicking on the icon

Come, I am taking a picture
visible in black and white

Now, the objective is to

bring color back to our
photos... For that, I will
use the Process /Debayer
In this window, there is
some elements to change
for the process to work-
Make and drain the...
set of our photos...

The sensor is a CMOS

therefore with a matrix
of type RVVB (2 times green).
To do this, we modify "Bayer/
mosaic pattern in "RGGB".

We are not going to de-bayouize.

only one photo but
all the photos. For this,
I click on 'Add Files' in
Target Images

Finally, I will indicate a

directory of witches
de-bayerized photos in
adding a prefix or suffix
in the name of the file (opponent)
in 'Output Files'.

Then we can launch the pro-

access by clicking on:
 ALIGN THE PHOTOS
Now that the photos
they are all in color, we
can move to the next step of
the alignment.

To do this, we will use the

processStarAlignment

Some elements are to

fill in like:

"Reference image": it is
the reference image and the
best of your series of
photos...

«Target Images» : l’ensemble

photos to align.

In the 'Output' tab

Images", it indicates the re-
folder where the photos will be
deposited after alignment.
Il n’y a pas de configura on
standard because it depends
from the photo but we can
"play" on some re-
settings to optimize the ali-
gnement.

"Detection scales" : I keep

the value of 5.

"Noise scales": I indicate

a value of 1. The more we
the higher it increases, the more the pixels of

noise will not be considered

dreams like stars.

Hot pixel removal: I am-

dyke 1 as value. I do not
set to 0 because I do not want
that a hot pixel is considered
to enhance like a star...

"Log sensitivity": I indicate

-1.50. The lower the number goes,
the more I want to detect of pe-
Your stars.
I can start the alignment by clicking on:
«Descriptor type» : Triangle
similarity.

STACK THE PHOTOS

Now, we are going to em-

pile all the photos
in order to form just one!

I will use the process /

Image Integration

I am going to add my photos

aligned by clicking on 'Add
Files.
I complete the 'Image' tab
Integration

Combination: Average
Normalization: Additive with
scaling
Generate integrated image
Evaluate noise
Automatic buffer sizes

I complete the "Pixel" tab.

Rejection (1):

Rejection algorithm depends

of the number of photos:
Sigma clipping or Ave-
"Raged Sigma Clipping" for
10 or 20 photos (to be tested on
your photos)
Winsorized Sigma Clip-
ping" for more photos..

I particularly check 'Clip

low pixels, clip high pixels
clip low range, Map range
rejection...

I am completing the 'Pixel' tab.

Rejection (2)

Sigma low: 5
"Sigma high": 3

I start the stacking in

clicking on:
I am going to get a photo
very dark. To see better
le résultat de l’empilement
some photos, I will use
the processScreenTransfer-
Functionand click on:

My photo appears well in

colors but with a strong
dominante verte. C’est nor-
bad...

My image is poorly centered.

(bad framing) so I
you will not be deleted alone
the borders of my
image because they are very dark-
but also a large
from the upper zone to
left

To crop my photo, I
are you going to use the process/Dyna-
micCrop

It is enough to draw the new one

rectangle on the photo in order to
crop the image.

To apply the crop, it

just click on:
I am going to give to my photo
a more natural color
and in a linear way with the
processScreenTransfer-
Functionand click on:

then

I therefore obtain an image

with colors quite equi-
freed (filter IDAS NB1).

Upon closer inspection, I

observed in the depths of the sky
a gradient and a red for-
statement present on the right
from the photo...

To reduce or even eliminate

my ugly gradient,
I will use the process/Dy-
Dynamic Background Extraction
I always use this process
DBE for the nebulas because
we can place the points of
manually referenced.

Note: The process/Automa-

Background Extraction
correspond plus à un traite-
mind with galaxies by
example...

The DBE window is quite

simple. Some points are
to be modified according to the photo:

«Default sample radius» :

corresponds to the size of the
decommissioning zones...
If I have a density in stars
strong then I reduce the size.

«Sample per row» : nombre

maximum of overlapping zones
Allons par ligne. We indicate
the number then we click
on "Generate". We obtain
dots on the image that
one can move simply
movement...
Note: I prefer to place my-
annually all my own
areas so do not click on
"Generate"...

«Sample color» : couleur

sampling areas,
I indicate a 'pink' well.
visible...

Correction: Subtraction
I check "Normalize" and "Re-
place target image
One can clearly see the areas
of samples that I have
indicated in the photo. They
are marked by small
pink squares...

To start the process, I

click on:

A large part of the gradient

has disappeared but remains in-
create a slight red zone
on the right side of the image...

I renew the process once

second time by clicking on:

My gradient is almost-
completely erased. If
need, I renew the opera-
where I apply an ABE in
pretending not to
losing signal... In my
case, I apply an ABE for
finalize everything.
Now, I will return
the more neutral background of the sky,
that is to say the closest
a neutral gray. For that,
we equalize automatically
the color curves.

We use the process / Color-

CalibrationBackground-
Neutralization

Two important points to

specify in this window:

Reference image
responds to a preview of
your sky background in the
photo. To do this, we create a
"preview", we are looking in
our photo the most
dark in the background of the sky...
To create a 'preview', we
click on this icon:

Then, we indicate it in 'Refe-

reference image.

Then, we determine the

value "Upper limit". With
the mouse pointer, we
point a dark area of
background of the sky and we read the value
the highest at the bottom of the
software... We will then take the
value greater than the maximum
let it be 0.09

We apply this process in

clicking on :
The obtained image is very
dark or colorful. It suffices
to click in the process /
ScreenTransferFunctionon
the icon:

then

The difference is slight but

very real. Our sky background
is more neutral on average.

Photo BEFORE / AFTER

After returning the background of

neutral sky, we are going cali-
Blur the colors of the photo.
For this, we use the process
Color CalibrationColorCa-
libra on

For the calibration of

colors are realized, it is necessary
indicate 2 reference images
referring to Pixinsight.

White reference :
"Reference image" is a
"preview" of a clear area
from the photo using
the icon

Background Reference :
"Reference image" is
a "preview" of an area
darkness of the background sky in
using the icon:

I also check 'Struc-

true Detection.
In the picture opposite, we say-
Here are the 2 previews:
- «Preview 1» : Zone claire de
the nebula
Previex 2: Dark Zone
from the bottom of the sky.

We start the process in cli-

how about the icon:

In the image on the right, we

you can clearly see the interest of the process

Color Calibration

A person who helps me sometimes to

remove a star or a
default... It's the process /
CloneStamp

Just press Ctrl + click

left mouse button for selec-
Select the area to copy. Then,
we place the mouse where it will be
apply the "paste".
The star has been successfully erased.
in this specific case.

Now, I am going to realize

an important step is
the histogram rise. The
but being of spreading the light
in many nuances
in order to reveal the most
dim lights. For that,
where we use our photo
we are going to turn off the light
treated linearly.

We use the process /Screen-

Transfer Functionwhere one
click on:

Before undertaking the climb

manual histogram
However, I use the process/
AutoHistogram

In this window, I will

just indicate "Stretch me-"
Logarithm Transform
moon.

Can I start the process:

I am obtaining an image with a
weak visualization of my
subject. I am done with this
process.

I am going to make a climb

of the histogram in a way
manual with the process /
Histogram Transformation

Thanks to the previous process,

I am making a nice graph.
legible and not glued
on the left. Thus, it is
easier to manage the color
black.

The principle is quite simple.

I am going to take the cursor
the black and position it to
the base of my curve side
left

With the gray tone slider

(the one in the middle), I'm going to also-
element the approach to the base
from the curve on the right side (without
too close to it all
even).

Can I start the process:

I am getting an image more
claire avec de nombreuses
nuances. The graph is
also wider.

In general, I repeat this.

operation 1 or 2 times. This
remains to the appreciation of
each one but it should not
that the object is neither too clear nor the
background of the sky...
For this, I set the value to 0.
previous settings in cli-
about

After a 3rd climb

histogram, I obtain this
image...

Next step: saturate the

colors of the nebula.
To do this, we can easily create
create a mask of the object
on Photoshop. On Pixinsight,
the creation of the object mask
it is difficult to achieve so little
of contrast and an image
noisy.

First, I will export my

Pixinsight photo in TIFF
exploitable sous Photoshop.

In the format, I select-

16-bit signal.
I open the TIFF file in
Photoshop. I create a new-
veal overlay: Overlay / New-
calf / Overlay

I am adding a new layer

above the photo.

I select the Lasso tool.

I will draw with the lasso

the approximate outline of
the nebula. I obtain a
selection in bullet points.

As it will be a lack,
I will fill this area.
of white: Edition / Fill /
Fill with white.

Now, I am going to inv-

be the selection: Selection /
Intervene.

I will fill this area with

black: Edition / Fill / Rem-
play with black.
To avoid a contrast
of strong top treatment, we
apply a Gaussian blur on
the whole image
Filter / Blur / Gaussian Blur

For a smooth transition,

a radius of 100 is indicated
pixels.

I am saving the obtained image

and I will open it in Pixin-
sight.

We find on Pixinsight
the photo of the nebula and
the image of the mask.

Reminder: the white zone

the mask will be able to be treated
driven by processes while
the black zone is protected.

To apply the makeup,

I will drag the tab
just the mask object
under the photo tab.
Thus the mask is laid down.
To verify it, the tab of the
photo changes color...

I am attaching my photo with my

red mask.

If we don't see the mask,

c’est que la visualisa on n’est
not activated. In this case, we
the egg in the "Mask" menu
Show Mask.
To see the saturation well
the colors of the nebula,
we make the mask invisible
by clicking on:
Mask / Show Mask

We are going to open the process

Color Saturation

In this window, one can:

either saturate the entire

colors while raising the bar
colors (you need to selec-
give the point to the left of
the line and the rise so that
the whole line goes up).

either place points on the

line by clicking on it at
places we want to saturate
specific colors. Then he
We need to raise the targeted points.

To initiate the process, we

click on the icon:

We repeat the opera

many times as long as we bear it
yet always watching
that the color does not saturate
and that the image be the most
natural possible...
Next step: strengthen the
born of the nebula. For
for that, we use the process /
UnsharpMask

This process is interesting

to reveal some de-
tails of a nebula by
example. On the other hand, to force
on this process will lead
a lot of noise.

For that, we use unique-

the 2nd slider that we
place everything on the left and we
choose for example the value
from 0.30.

To start the process, we

click on the icon:

One can repeat the opera

several times as long as the
sound remains almost imperceptible
ble.

Next step: make the

the smoothest background sky
possible or the least noisy
without touching the object or the
stars. As seen prec-
crazy, we already have
a mask of the object here
against.

We are going to make a mask

of stars in the photo. In
reality, often we do it at
minus 2 because Pixinsight doesn't know
not "capture" all sizes
of stars at the same time...
First, we will reduce the
object mask window in
clicking on:

To create a mask
of stars, we will use the
processStarMask

Let's start a first

mask for little ones
stars.

In general, I keep these screams-

terms for the photos:
Scale : 5
Large-scale : 2
Small-scale : 1
Compensa on : 2

Smoothness : 12

The important criterion is

the management of the stars by
report on noise therefore in
in my case I am at 0.40 for the
"Noise Threshold" for a
correct result.
The noisier the image, the more
you will need to move this slider
to the right.

We are starting the process in CLI-

how about:

Note: to go faster, I
can create a preview-
by clicking on the icon:

Here is my pet mask

stars... I click 2 times on
the tab of her name and I
rename "Pe tes_etoiles"
I reduce this mask of pe-
Rate the stars by clicking on:

Now, we are making the

2nd mask for the
average stars.

In the StarMax process,

he suffers from falling towards the
right quite far the cursor
"Midtones" either for me
photo 0.97 (but never 1).

This will allow to select-

Give the stars size
average.

We are starting the process in CLI.

how about:

So we are awaiting a 2nd one

mask with the stars
averages.

I rename this mask to

Average_stars
We have 3 masks:
Mask of the object
Mask of little stars
Average star mask

We will group these

3 masks to make them
that one! To do this, we use-
we will be the processPixelMath

We open the window of Pixel-

Math and click on 'Ex-
Editor’s pressure. We slide
a first mask towards the
left window AND between
for each mask, we put a
«+».
We validate with an 'OK'.

Under the 'Desgnation' tab,

we check 'Create new'
image.

We start the process in CLI.

how about:

We are now getting a

unique mask!

We are going to apply this

mask in the photo.
We place this mask on the
photo. ATTENTION: The red
signifies the protected area or
we wish on the contrary
soften the background of the sky...

In that case, it is enough to in-

pour the mask:
Mask / Invert Mask

To facilitate processing of the

sky background, it is preferable
not to see this 'red'.
To do this, we will hide the
masque : Mask / Show Mask

The noise reduction of the sky background

can start. We will use
the process /TGVDenoise

Either I work with a

"preview" either I zoom in on
the background of the sky.

In the TGVDenoise window,

I move to the right the
cursor "Strength". In this
here, we are around 8.

Can I launch 1 or several?

repeat the process by clicking:
This process allowed for smoothing
our background sky all in
keeping the details of the ne-
blues singer.

Photo on the side showing the

sky background before AND after
treatment...

The photo is quite okay.

To further improve it, we
let's calibrate the color by
photometry. For this, we
let's use the process/Photo-
metric Color Calibration

This eye is powerful because it

go correct our mistakes of
colorimetry in photography.
For this, a few fields to
fill in:

The coordinates of the image

to be filled in manually with
the site [Link] or in
automatically by clicking on
Search Coordinates.

The focal length (in mm) and the size

the pixel of the sensor (in µm)

Check "Force plate ground-

living

We start the process in CLI-

amount on:

We obtain a series of gra-

physics on the modifications
colorimetric...
Thanks to this process, we
getting a photo as soon as possible
close to the color stan-
pain, which allows a cor-
color recognition objective.

With Pixinsight, it is pos-

able to correct small errors
errors of opaqueness... Thus
We will have more details.
fins and stars no longer born.
We use the process /Decon-
volume

Pour que cela fonc onne, il

it is necessary to create a
average star of our
photo in order to correct everything
the photo. We use the process
/ DynamicPSF

Once the window is open,

I will click on my photo
on about twenty (or more)
well-repaired stars without
choose those that are close to the
board.

In the right table,

each selected star
will present its specificities...
If a star shows a
Unusual FWHM, it needs to be
delete by clicking on it
line then pressing on:
Once the selection is complete
Well, we're going to create a star.
average synthesis.

For that, we click on:

And I am getting my PSF of re-

reference. I can close the
DynamicPSF window.

I will open the process /

Deconvolution

A few points to accomplish:

We click on 'External'
PSF» and we load the photo
PSF completed just before.

We can play on an 'Ite-

very low reasons for a
smoother effect.

- «Target» : «Luminance»
even if it is a photo
colors.

I check 'Deringing' and

I push the cursor a little.
"Global dark" (here 0.20) aim
de supprimer des artefacts.

I can start the process in

clicking on:
The process brings in name-
breux advantages: stars
rounder, thinner, and more
details that are...
violent in the nebula...
The function correction!

BEFORE / AFTER

In this photo, we have

many stars and
to enhance the value of me
nebula, we are going to do
kill a reduction of stars.

We begin by realizing
a mask of stars with the
processStarNet

Before using this process, it

You need to install Starnet via this.
page : h ps://sourceforge.
net/projects/starnet/

We get a directory with

of files...

We are opening the StarNet process

and we will just indicate the
file access paths
.pb

I check "Create star Mask"

to obtain the mask
of stars...
To get my mask, I
start the process:

I am putting on the mask (to the right)

and my photo without stars (to
left). For this last one,
I want to keep my stars
So I select my photo
and I cancel the last action
by clicking on:

To make the processing

sweeter and more natural, I will
increase the perimeter of
stars in the mask in
using the processHisto-
gram transformation

In this process, I will just

take the gray cursor,
gently push it on the
left, and slightly bumpy
on the curve. Then I apply
by clicking on:

I am getting more stars

larges...
I slide the mask tab
just below the tab
of the photo. Immediately,
I see that the background of the sky and
the nebulae are protected.

I make the mask invisible

to better visualize the trait
next statement in Mask /
Show Mask

Now, we move on
on the reduction of stars in u-
reading the processMorpholo-
gicalTransforma on

In this window, a few

points :

«Operator» : Morphological
Selection

"Interlacing" and "Iterations":

1 (to go gently)

«Amount» : 0.50 (un traite-

don't speak too loudly)

"Selection": 0.25 (for a

gentle erosion of the star

«Size» : 5 et je clique que la

shape of a star:

I am starting the process:

We can see the stars well
the finer I mean better in
value the object...

We are going to annotate our

photo thanks to plate solving
dans Pixinsight.

We are going to use the Script /

Image Analysis /ImageSolver

A few clarifications for that

it works:

It is necessary to indicate the coordin-

born from the subject of the image.
For this, I use the button
Search

I also indicate the focal length

of my instrument and the size
of a pixel from my camera.

I launch the script by clicking

Okay.
Once the script is finished, it
it is possible to annotate the pho-
So, for that, I will use the
Script / Render / AnnotateI-
image

Just click on this

that we wish to make appear
to shine in the annotated photo
NGC, Messier, stars...

For better visibility,

I change the colors of the
text.

I start the script by clicking

OK.

I take the photo with the an-

desired notes.

Thanks to plate solving, the

the header of our file is
updated. We can see it.
with the ProcessFITSHeader
 SUBJECT: NGC 281
CAMERA : ATIK 460EX Mono
60 x 60s

FILTER: BAADER Ha 7nm

WITHOUT SELF-GUIDANCE
WITHOUT DOF
The photo is taken by a ca-
my monochrome CCD with
a Ha filter. It is the nebula-
use Pacman or NGC281.

We will resume the treaty-

by following various
steps to 'improve'
our image.

I open the photo on Pixin-

sight. It is very dark, it's
normal, she was taken with
a Ha filter...

I will brighten the photo in

linear with the process /
ScreenTransferFunction

I click on the icon:

I note following the stacking-

while the borders of
the photos are irregular,
sometimes too dark... I
you will crop the photo with the
processDynamicCrop

To apply the crop, it

just click on:
The background of the sky seems quite
coherent and very little gra-
I'm just going to apply.
the processAutomaticBack-
ground extraction

I apply the values by

default except "correction":
Subtraction

I am starting the process:

J’ob ens un fond de ciel sans

no gradient.

Now, I am going to realize

an important step, it is
the histogram rise. The
but being to spread the light
in many shades
in order to reveal the most
dim lights. For that,
we use our photo where
we are going to turn off the light
linearly processed.

We use the process /Screen-

Transfer Functionwhere one
click on:

The photo becomes complete again.

black mind and it’s normal.
Before making the climb
manual histogram
Well, I use the process /
AutoHistogram

In this window, I will

just say 'Stretch me-'
Logarithm Transfor-
maon.

Can I launch the process:

The image is clearer but

we can only see some
stars... The goal of this der-
the process is to facilitate
the extension of the curve of
the following histogram.

Now, we will
raise the histogram in order to
to make all appear
elements of the photo with the
processHistogramTransfor-
ma on

Obviously before using

this process, it must be uncovered
the processScreenTransfer-
Functionwhere one clicks on:

In the windowHisto-
gram transformationfor-
visualize the curve on click
on:

and to see the changes

live:
We will first bring closer
the black cursor (the one
to the left) towards the base of
the curve that darkens the
photo.

Similarly, we bring closer

gently the cursor of
white (on the right) of the base
from the curve that illuminates the
photo...

We see the nebula appearing

reading on the preview.
To apply it to the photo,
we slide this icon on the
photo :

The process is well applied.

in the photo. Our window
of preview must be
reset by clicking on:

We will reiterate this pro-

cess to spread the light.

This step can be accomplished.

2 or twice until obtaining a
correct image...
I thus obtain a photo
with a good balance and a
ample histogram.

We can improve the

contrast between the object and
the sky background. For that,
we use the processCur-
Transformation

Forto visualize the curve

we click on:

and to see the modifications

live:

On the curve, we place a

point corresponding to black
from the photo and a white dot
corresponding to the area
light of the object.

Then we play on these 2 points

to form an "S". Thus,
we darken the background of the sky
and we lighten the whites.
Attention not to overload the
white...

We get a photo with

a better contrast.

We apply on the photo in

clicking on:
We are going to correct lightly-
with errors of optics
with the processDeconvolu-
on

For it to work, it
it is necessary to create a
average star of our
photo in order to correct everything
the photo. We use the process
/ DynamicPSF

Once the window is opened,

I am going to click on my photo
on about twenty (or more)
of well-repaired stars without
choose those that are close to the
border.

In the table on the right,

each selected star
will present its specificities...
If a star presents a
Abnormal FWHM, it is necessary to have it.

delete by clicking on it
line then by pressing on:

Once the selection is complete

No, we are going to create a star.
average synthetic

To do this, we click on:

And I get my PSF back

reference. I can close the
DynamicPSF window.
I will open the process /
Deconvolution

A few points to achieve:

We click on 'External'
PSF» and we load the photo
PSF completed just before.

In the case of this nebula-

leuse, I apply an "Itera-
very low (2) for a
softer effect.

I check 'Deringing' and

I'm pushing the cursor a bit.
Global dark (here 0.20) for
to remove artifacts.

I can start the process in

clicking on :

I am receiving an image with

rounder stars, more
presents and details in the
nebula.

A final step (for

this photo) will be de-
louder. For that, we use
the process /MultiscaleLi-
nearTransform

The advantage of this process is

so that we can make a mask
in order to preserve a little the
nebula but to apply
a stronger treatment on
the background of the sky by protecting
the stars.

To create the mask, we

car "Linear Mask" and "Pre-
view Mask" then we click
on the preview icon:
For the mask and for this
precise photo, we intervene
on the slider 'Amplify-
on» (here 1 in order to see in
light black the nebula and a
white sky background.

We add 'Smoothness'
(here 3) to avoid arte-
facts and especially soften them
transi ons...

Once that is done, I uncheck.

"Linear mask" and I review
my photo in preview.
I am closing this window.
preview and we obtain
this screen next to.

I will indicate "Layers": 7

"Detail layer" and "Noise Re-

"duc on" must be checked.

According to the strength of the smoothing, I

starts on 'Iterations'
1

According to the desired outcome, I

reiterates several times the pro-
access by clicking:
SUBJECT: Group M66
Photo of Sébas in Tiersoone

CAMERA: ASI 294MC

20 x 300s

FILTRE : SANS FILTRE

WITH AUTOGUIDANCE
WITHOUT DOF
To obtain a final photo
very correct, it is important
to select the best
their photos from the shooting.
To do this, I will carry out a
first visual selection
with the PROCESSBlink

This owl is a kind of

photo viewer... For
select my photos, I
click on:

Once the photos are integrated,

it is enough to use the arrows
to verify manually
if a photo contains
traces of satellites or a
cloudy passage

For a large number of

photos, I can take a
automatic reading all
half seconds and throw
the soul

If I find a photo that

I do not wish to keep,
I can move it in a
another directory (for example
«trash can») by clicking on :

Here I notice visually

some photos upside down
due to the turnaround at
meridian...
After a first selection
in a visual, it is possible to
select the photos of fa-
one more rational thanks to the
ProcessSubframeSelector

Images need to be added to

study by clicking on:

First, to measure our

photos, we choose:
Rou ne: Measure Subframes

Then, it must be indicated

some information
notably in 'System
Parameters at the level of
"Subframe scale" that one
can calculate here: h ps://
[Link]/echan llon-
swimming-and-visual-field/

It is enough to enter the focal length of

the instrument (mm) and the size
of a pixel (µm).

Then, we can start the

process by clicking on:
After a few seconds
of analyses (according to the number
from photos), we obtain a
new window with a
graph...

In the table, I have the list

of my photos. In the
graph below, I have the
measurement of the FWHM of my
photos. The lower it is,
the more the stars will be born
and pee.

This is where I select

the photo with the FWHM there
lower because it will serve me
reference photo for
the alignment...

Now, I am going to delete-

see the photos where the FWHM
is bad and therefore where it
is the highest. (3 photos)
For this, at the level of the gra-
just click on
the points (corresponding
aux photos) que je ne sou-
don't keep the deal
It is noted that
the points turn into
cross...

Finally, I will be able to save-

keep the 'good' photos.
I'm switching to Rou ne mode:
Output Subframes.

In the 'Output Files' area,

I indicate the directory where
my 'good' photos will be
Then, I validate in
clicking on :
I am going to open my photo of
reference on Pixinsight, that one
with a very small FWHM.
She appears very black and
it's normal.

So that my photo is
visible, I will use the
ProcessScreenTransfer-
Function

This or that simply allows

to visualize the photo without
modify the pixels.

By clicking on the icon su-

Vante, I obtain an image
visible in black and white:

Now, the objective is to

give color back to our
photos... For that, I will
use the ProcessDebayer
In this window, there is
a few elements to change
to ensure that the process works
It is worth it and debayerizes the ...

set of our photos...

The sensor is a CMOS coupled

their therefore with a matrix
of type RVVB (2 times green).
For this, we modify 'Bayer/
mosaic pattern in RGGB.

We are not going to de-bayouize

only one photo but
all the photos. For that,
I click on 'Add Files' in
Target Images

Finally, I will indicate a

directory of sources of these
de-bayerized photos in
adding a prefix or suffix
in the name of the file (opponent)
in 'Output Files'.

Then we can launch the pro-

access by clicking on:
Now that the photos
are all in colors, we
can move on to the step of
the alignment.

For this, we will use the

processStar Alignment

Some elements are to

fill in like:

Reference image: it is
the reference image and the
best of your series of
photos... (here photo 16).

«Target Images» : l’ensemble

photos to align.

In the 'Output' tab

Images", we indicate the re-
folder where the photos will be
deposited after alignment.
There is no configuration
standard because it depends
from the photo but we can
play on some re-
adjustments to optimize the ali-
management.

"Detection scales": I keep

the value of 5.

"Noise scales": I indicate

a value of 1. The more we
the increases, more the pixels of
noise will not be considered
dreams like stars.

Hot pixel removal

dam 1 as a value. I do not
set to 0 because I don't want it
that a hot pixel is considered
spotted like a star...

"Log sensitivity": I indicate

-1.50. The lower the figure goes,
the more I want to detect from pe-
The stars.
I can start the alignment by clicking on:
«Descriptor type» : Triangle
similarity. U le for the re-
tournaments at the meridian.

Now, we are going to em-

pile the entire set of photos
aligned in order not to form
What a!

I will use the process /

Image Integration

I am going to add my photos

aligned by clicking on "Add"
Files.
I am completing the 'Image' tab
Integration

Combination: Average
Normalization: Additive with
scaling
Generate integrated image
Evaluate noise
Automatic buffer sizes

I am completing the 'Pixel' tab.

Rejection (1):

Rejection algorithm depends

of the number of photos:
Sigma clipping or Ave-
"Raged Sigma Clipping" for
10 or 20 photos (to be tested on
your photos)
Winsorized Sigma Clip-
ping" for more photos..

I especially check 'Clip

low pixels, clip high pixels,
clip low range, Map range
rejection...

I complete the "Pixel" tab

Dictionary (2):

"Sigma low": 5
"Sigma high": 3

I start the stacking in

clicking on:
I am going to get 3 very photos.
shadows. I keep unique-
see the photo "Integration".
To see the result better
from the stacking of the photos,
I will use the process/
Screen Transfer Functionand
click on:

My photo looks good in

colors but with a strong
dominant green. It is nor-
bad...

My image is poorly centered.

(bad framing) so I
you are going to delete not only
my borders of
image because they are very
dark.

To crop my photo, I
are you going to use the process/Dyna-
micCrop

It is enough to draw the new one

rectangle on the photo in order to
crop the image.

To apply the crop, it

just click on:
I will give to my photo
a more natural color
and in a linear way with the
processScreenTransfer-
Functionand click on:

then

I therefore obtain an image

with weak colors.

NOTE: Having no Flats,

one observes a vineyard
important and donuts.

To eliminate these defects

opques, I will again
crop the photo unfortunately
cement...

A slight gradient is present

sur la photo, je vais u liser
simply the processCar-
macBackgroundExtractor
Note: The process/Automa-
Background Extraction
corresponds more to a trait-
statement with galaxies
unlike the DBE. But
sometimes we use both pro-
cess...

The ABE window is quite

simple. A few points are
to be modified according to the photo:

Correction: Subtraction

I check 'Normalize' and 'Re-

place target image.

I am starting the process in CLI.

when on:

Thanks to the process, I obtain

a perfectly sky background
homogeneous.
Now, I am going to give back
the more neutral background of the sky,
that is to say the closest
in a neutral gray. To do this,
we equalize automatically the
color curves.

We use the process / Color-

CalibrationBackground
Neutralization

Two important points to

specify in this window:

"Reference image" corres-

respond to a "preview" of
your background sky in the
photo. To do this, we create a
"preview", we search in
our photo the most
dark in the background of the sky...
To create a 'preview', we
click on this icon:

Then, we indicate it in 'Refe-

reference image.

Then, we determine the

value 'Upper limit'. With
the mouse pointer, we
point to a dark area of
sky background and we read the value
the highest at the bottom of
software... We will then take the
value greater than the maximum
it is 0.09

This process is applied in

clicking on:
The obtained image is very
dark or colorful. That's enough.
to click in the process /
ScreenTransferFunctionon
the icon:

then

The difference is small but

very real. Our sky background
is more neutral on average.

After returning the background of

neutral sky, we are going to cali-
bring the colors of the photo.
For this, we use the process.
Color CalibrationColorCa-
libra on

For the calibration of

colors are realized, one must
indicate 2 reference images
reference to Pixinsight.

White reference :
"Reference image" is
a "preview" of an area
shadow from the depths of the sky in
using the icon:

Background Reference :
"Reference image" is a
"preview" of a clear area
colored by a galaxy or
of a nebula using
the icon:

I also check "Struc-

True Detection
In the image on the opposite side, we say-
Read well the 2 "previews":
Preview 1: Dark zone
from the sky
- "Preview 2": Light zone
of a galaxy

On lance le process en cli-

how about the icon:

We obtain a nearly photo-

black hand. On the process /
ScreenTransferFunctionon
then click on:

then

We get a fairly image

balanced.

We are going to discover the pro-

cessScreenTransferFunction
by clicking on:
Before making the ascent
manual histogram
well, I use the process /
AutoHistogram

In this window, I will

just indicate 'Stretch me-
Logarithm Transform
ma on.

Can I start the process:

I am taking an image with a

weak visualization of my
object. I have finished with this
process.

I am going to make a climb

of the histogram in a way
manual with the process /
Histogram Transformation

Thanks to the previous process,

I am making a good chart.
readable and not stuck
on the left. Thus, it is
easier to manage the color
black.

To see the curve, I click

on:
The principle is quite simple.
I am going to take the cursor
of the black and the positioning to
the base of my curve side
left

With the gray tone slider

(the one in the middle), I am also going to
element approach to the base
from the curve on the right side (without
too close to it all
same).

Can I start the process:

I obtain a clearer image

claire with many
nuances. The graph is
also wider.

Here, I repeat only once

this operation.

This remains to the appreciation of

each one but it should not
that the object is not too clear nor the
sky background...
For this, I set to 0 the
previous settings in cli-
how much on:
Now we are going to do
appear details in
the very white areas of
galaxies.

To do this, we will create a

mask with the process /
Range Selection

By default, we select 'In-

green." We are progressively increasing
Upper limit
only see the galaxies. From
even, we increase 'Smoo-
thness" to make disappear
the stars and soften the
transitions.

We are obtaining this image in

clicking on :

This image will become our

mask: the white will be modified
the protected black.

We slide the mask tab

under the tab of
our photo.

The mask is red.

To make it invisible but

When selecting in the
menu Mask / Show Mask
To give more details,
we use the processRange
Selection

I indicate in this case:

«Number of layers» : 6
Number of iterations: 1

I mainly check "Deringing"

I am starting the process:

The image shows a lot of

details.

Protons from the mask on the

galaxies... We will again
make the galaxies clearer
by using the process /
Unsharp Mask

The photo being slightly

noisy, I will increase by
very little the "StdDev" to a
value of 0.90.

We apply by clicking on:

Likewise, we are going to satu-
see the color better
make stand out...

We are going to open the process

Color Saturation

In this window, one can:

either saturate the entire

colors while raising the bar
colors (it is necessary to select-
give the point to the left of
the line and the ascent so that
the whole line goes up).

be it placing points on the

line by clicking on it at
places where we want to saturate
specific colors (here the
brown and blue). Then it is necessary to
raise the targeted points.

To start the process, we

click on the icon:

They are repeating the opera

several times as long as we can
assistant and always monitoring
so that the color does not saturate
and that the image be the most
natural possible...
We will soften the background
from heaven. To do this, we are going to
reverse the mask (it was
applied to the galaxies) by
Mask / Invert Mask

In red, the galaxies are

protected.

We make the layer invisible.

either Mask / Show Mask

To denoise the background of

Sky, we use the Process /
MultiscaleLinearTransform

Some information:

«Layers» : 6

"Noise Reduction": to check

and I apply the value by
default for each layer

We apply by clicking:
I obtain a background of sky with
very little noise.

The photo is rather correct.

To further improve it, we
let's calibrate the color by
photometry. For this, we
let's use the process / Photo
metricColorCalibration

This owl is powerful because it

go correct our mistakes of
colorimetry on the photo.
For this, a few fields to
fill in:

The coordinates of the image

to be filled manually with
the site [Link] or aen
automatic by clicking on
Search Coordinates

The focal length (in mm) and the size

the pixel of the sensor (in µm)

Check 'Force plate ground

living

We launch the process in CLI-

about

We obtain a series of gra-

physics on the modifications
colorimetric...
The photo is quite good.
To further improve it, we
let's calibrate the color by
photometry. For this, we
let's use the processPhoto-
metric Color Calibration

This wolf is powerful because it

go correct our mistakes of
colorimetry in photography.
For this, a few fields to
fill in

The coordinates of the image

to be filled in manually with
the site [Link] or in
automatic by clicking on
Search Coordinates.

The focal length (in mm) and the size

the pixel of the sensor (in µm)

Check 'Force plate ground-

ving

We launch the process in CLI-

how about:

We obtain a series of
graphics with the modifi-
colorimetric charts...

You get a photo with

calibrated colors for
the galaxies and a background of sky
darker.
 SUBJECT: Iris Nebula - NGC 7023
Pre-processing & treatment

CAMERA : ATIK 460EX 40 x 180s

FILTER: IDAS NBZ (filter not suitable for this target)

WITH AUTOGUIDING

WITH DARK, FLAT and DARK FLAT

 CASE STUDY: BLACK AND WHITE PHOTO WITH DARK and FLATS, DF
To obtain a final photo
very correct, it is important
to select the best
their photos from the shooting.
To do this, I will carry out a
first visual selection
with the oul PROCESS /Blink

This owl is a kind of

photo viewer... For
select my photos, I
click on:

Once the photos are integrated,

it is enough to use the arrows
to check manually
if a photo contains
satellite tracks or a
cloudy passage

For a large number of

photos, I can do a
automatic lecture all
the half seconds and throw
the animal

If I find a photo that

I do not wish to keep,
I can move it in a
another directory (for example
"trash can") by clicking on:
 SELECT THE BEST PHOTOS
After a first selection
visually, it is possible to
select the photos of fa-
more rational thanks to the
Process /Subframe Selector

We need to add the images to

study by clicking on:

First, to measure our

photos, we choose:
Rou ne: Measure Subframes

Then, it is necessary to indicate

some information
notably in 'System
Parameters at the level of
«Subframe scale» qu’on
can calculate here: h ps://
[Link]/echan llon-
swimming-and-visual-field/

It is enough to enter the focal length of

the instrument (mm) and the size
of a pixel (µm).

Then, we can launch the

process by clicking on:
After a few seconds
of analyses (depending on the number
from photos), we obtain a
new window with a
graphic...

In the table, I have the list

of my photos. In the
graph below, I have the
measurement of the FWHM of my
photos. The lower it is,
the more the stars will be born
and pets.

This is where I select

the photo with the FWHM there
lower because it will serve me
reference photo for
the alignment... (Here it is the
photo 6).

Now, I will delete-

but the photos where the FWHM
is bad and therefore where she
is the highest.
For this, at the level of the gra-
it is enough to click on
the points (corresponding
from the photos) that I do not-
don't keep to the treaty
It will be noted that
the points transform into
cross...

Finally, I will be able to save-

keep the 'good' photos.
I'm going into Roun mode:
Output Subframes.

In the 'Output Files' area,

I indicate the directory where
my 'good' photos will be
submitted. Then, I validate in
clicking on:
 PREPROCESSING: CREATE YOUR DARK MASTER
Our photos were taken
in 180s of pause. This created
the noise. It is easy to see on
a photo, we see dots
saturated whites. To avoid this
thermal noise, I realized
the Darks. Here, I made one of them.
ten

We are going to create a Master's program.

Dark from the whole

the Darks.

For this, we will use the

ProcessImage Integration

We add the Darks files

by clicking on "Add Files".
It is important to fill in the
criteria as shown opposite in order to
to obtain the best Dark
possible.

We click on this icon

to initiate the process:

We have 4 images
shadows. We are going to keep
only the photo
the "integration".

We save this file with

the name "MasterDark".
 PREPROCESSING: CREATE YOUR MASTER DARK FLAT
To properly calibrate the Flats,
we need Dark
Flats.

We are going to compile here the

Dark Flats (20 photos) with the
ProcessImage Integration

We add the Darks files

Flats by clicking on "Add"
Files

We do verify the para-

meters in "Image Integra-
"on" and "Pixel Rejection 1".

We click on this icon

to launch the process:

We received 4 images.
sombres. Nous allons garder
only the photo
the "integration".

We save this file with

the name "MasterDarkFlat".
 PREPROCESSING: CALIBRATE THE FLATS

We are going to calibrate our Flats

with the MasterDarkFlat and the
MasterOffset.

We will use the Process

/ Image Calibration

Pour cela, on ajoute nos Flats

by clicking on "Add Files".

A directory is created at
name of 'CalibratedFlats'
Where will the Flats be located?
calibrated.

We will indicate where it is located.

our file 'MasterOffset'.

We will also indicate where

is our file 'Mas-
terDarkFlat».

Make sure to check the parameters

as in the picture.

We click on this icon

to initiate the process:

We therefore obtain a reaper-

toilet with all our Flats
calibrated...
 PREPROCESSING: CREATE YOUR MASTER FLAT

We are going to create the Master

Flat apart from Flat files
calibrated with the Process /
Image Integration

We add the files of

directory with the Flats
calibrated by clicking on "Add"
Files.

We check the para-

meters in 'Image Integra-
"on" and "Pixel Rejection 1."

We click on this icon

to initiate the process:

We obtain 4 images
shadows. We will keep
only the photo in background
the "integration".

We save this file with

the name "MasterFlat".
 PREPROCESSING: CALIBRATE THE LIGHTS

We will calibrate our

images Lights.

We are going to use the Process

/ Image Calibration

To do this, we add our

Lights by clicking on "Add"
Files

We create a directory at
name of "CalibratedLights"
where will the Lights be located
calibrated.

We will indicate where it is located.

our file "MasterDark".
Carefully check the settings
as in the image.

We also indicate our

file "MasterFlat".

We click on this icon

to start the process:

We therefore expect a reper-

bathroom with all our Lights
calibrated...
 PREPROCESSING: ALIGN THE LIGHTS

We are going now

align our Lights with the
ProcessStarAlignment

We choose a calibrated Light

with the lowest FWHM
Reference image

We verify the other parts.

meters.

We add all the files

of the directory containing the
Calibrated lights by clicking on
Add Files

In the parameter "Output

directory, we are going to create a
directory "AlignedLights"
where will the photos be
aligned.

In the 'Description' parameter

On type, I select
Triangles.

We click on this icon

to initiate the process:

One obeys in the re-

take all our photos well
aligned...
 PREPROCESSING: CREATE THE MASTER LIGHT

Now that our images

are aligned with each other,
we must stack all
the photos.

We are going to use the Process

/ Image Integration

For this, we add our

Aligned lights by clicking
on 'Add Files'.

In 'Image Integration', we
select the right fields.
In 'Pixel Rejection (1)', we
select the right fields.

In "Large-Scale Pixel Re-

on selection we choose the
good fields.

We click on this icon

to start the process:

We have 4 images.

It is absolutely necessary to register

remove the image 'Integration' because
it is on this one that we al-
we will work on the treatment.
 TREATMENT
The photo is taken by a
monochrome CCD camera
with an Idas NBZ filter (not
the idea for this target). It is
the Iris nebula.

We are going to resume the treatment-

by following various
steps to "improve"
our image.

I open the photo on Pixin-

sight. She is very black, it is
normal...

I will brighten the photo in

linear with the process /
ScreenTransferFunction

I click on the icon:

I notice following the stacking-

while the image is quite
homogeneous. A slight bor-
a darker one reveals itself
on the left... I'm going to crop-
for the photo with the process /
DynamicCrop

I cut out the photo with the

mouse... Then, I apply the
crop by clicking on:
The background of the sky seems quite
coherent and very little gra-
Okay. I'm just going to apply.
the processAutomaticBack-
ground extraction

I apply the values by

default except for 'correction':
Subtraction.

I am starting the process:

I obtain a sky background without

no gradient.

Now, I am going to realize

an important step, it is
the rise of the histogram. The
but being to spread the light
in many shades
in order to reveal the most
dim lights. For that,
we use our photo where
we are going to turn off the light
treated linearly.

We use the process /Screen-

Transfer Functionwhere one
click on :

The photo becomes complete again.

Black mind and that is normal.
Before making the ascent
manual histogram
I use the process
AutoHistogram

In this window, I will

just indicate 'Stretch me-'
method: Logarithm Transform
maon.

Can I start the process:

The image is clearer but

we only see some
stars... The goal of this der-
the process is to facilitate
the alignment of the curve of
the following histogram.

Now, we are going to

mount the histogram in order to
to make all appear
elements of the photo with the
processHistogramTransfor-
ma on

Obviously before using

this process, it needs to be uncovered
the processScreenTransfer-
Functionwhere you click on:

In the windowHisto-
gram transformationto pour
visualize the curve on click
on:

and to see the modifications

live:
First, we will bring closer
the black cursor (the one
from left) to the base of
the curve that darkens the
photo.

Similarly, we bring closer

gently the cursor of the
white (to the right) of the base
from the curve what enlightens the
photo...

We see the nebula appearing.

read on the preview.
To apply it to the photo,
we slide this icon on the
photo :

The process is well applied.

in the photo. Our window
the preview must be
reset by clicking on:

We will reiterate this pro-

this to spread the light.

We can achieve this step

2 or twice until obtaining a
correct image...
I take a photo like this
with a good balance and a
ample histogram.

We can improve the

contrast between the object and
the background of the sky. For that,
we use the processCur-
Transformation

To visualize the curve

we click on:

and to see the modifications

live:

On the curve, we place a

point corresponding to black
in the photo and a white dot
corresponding to the area
clear of the object.

Then we play on these 2 points

to form an 'S'. Thus,
we darken the background of the sky
and we lighten the whites.
Attention not to saturate the
white...

We get a photo with

a better contrast.

We apply to the photo in

clicking on:
We are going to correct slightly-
ment of op que errors
with the process /Deconvolu-
on

For it to work, it
it is necessary to create a
average star of our
photo in order to correct everything
the photo. We use the process
/ DynamicPSF

Once the window is open,

I will click on my photo
on about twenty (or more)
well-repaired stars without
choose those that are close to the
board.

In the table on the right,

each selected star
will present its specifics...
If a star displays a
Unusual FWHM, it is necessary to have it

delete by clicking on it
line then by pressing :

Once the selection is completed

Well, we are going to create a star
average synthetic

To do this, click on:

And I get my PSF back

reference. I can close the
DynamicPSF window.
I am going to open the process.
Deconvolution

A few points to consider:

We click on 'External'
PSF" and we load the photo
PSF carried out just before.

In the case of this nebula-

I apply an 'Itera-
very low (2) for a
smoother effect.

I check 'Deringing' and

I push the cursor a bit.
"Global dark" (here 0.20) for
to remove artifacts.

I can start the process in

clicking on:

I obtain an image with

rounder stars, more
presents and detail in the
nebula.

We are going to strengthen the

details of the nebula with
the processUnsharp Mask

We test the effect by leaving the

a rather low ‘StdDev’ level,
with an "Amount" of 0.50.

We apply the effect in cli-

about quant:

We get a clearer image.

detailed. It is especially important not to
don't force it...
Now, we will
denoise the photo. To do this,
first we will iso-
read the nebula and the stars.
The goal is to denoise.
mainly the background of
sky.

We are going to duplicate the photo.

in Image / Duplicate

On the duplicated photo, we are going to

use the processACDNR

In the section 'Lightness'

Mask
then we click on:

We move slightly the

sliders in order to obtain a
nebula and stars
dark but a background of sky
clear.

This white image will help us

serve as a mask.

I validate the white photo in

clicking on:
We will put on our mask
in our photo, it is enough to
slide the mask tab in
bottom of the tab of our
photo.

The red zones are the

zones that will not be affected by the
treatment.

One can make this invisible

mask in Mask / Show
Mask

Now, we can
reduce noise with the Pro-
cessTGVDenoise

We are increasing the strength of

denoising with 'Strength'.
It's the same for 'Smooth-
ness

For a gentle treatment and

by level, I advise to
put a 'Iterations' of
50.

We renew the process

until his satisfaction on
the image.
 OBJET :Nébuleuse de l’Iris NGC 7023
Quick preprocessing: Weighted Batch Preprocessing

CAMERA : ATIK 460EX 40 x 180s

FILTER: IDAS NBZ (filter not suitable for this target)

WITH AUTOGUIDANCE

WITH OFFSET, DARK, FLAT and DARK FLAT

 CASE STUDY: BLACK AND WHITE PHOTO WITH DOF, Dark Flats
AUTOMATIC PREPROCESSING: Weighted Batch Preprocessing

To save time, it
is sometimes useful to use a
automatic process after
to have selected the best
their photos. Let's take a look at
closer to the Script /Weighted
Batch Preprocessing

On the right side of the screen,

it is useless to check differently
rents parameters such as:

"CFA images": click here if

you took pictures in
colors

Save frame groups on exit

check this box in order to
save the files /
directories to the next or-
opening of the process.

Auto-detect best reference

: to check so that the
select script for the ali-
management the best report
FWHM / Eccentricity

«Output Directory» : Indi-

what the directory where you
you will find the result of this
script about your photos.
 The OFFSETs
We will first
add the offsets by clicking
on the 'Bias' tab
Images must be added to
process by clicking on "Add"
Bias.

Combination: Average

"Rejection": Linear Fit Clip-

ping

The DARKs
We will now add
to the darks by clicking on
the 'Darks' tab.
Images need to be added to
process by clicking on 'Add'
Darks.

Combination: Average

Rejection: Linear Fit Clip-

ping
 The FLATs

We will now add

get the cats by clicking on
the 'Flats' tab.
We need to add the images to
process by clicking on "Add"
Flats.

Combination: Average

Rejection: Linear Fit Clip-

ping

ATTENTION: do not forget

to check the box "Calibrate"
with at darks» car in the
next step we will
add Dark Flats...

The DARK FLATs

To calibrate correctly
In our Flats, we have realized
the darks of ats... Other-
said, they are Darks.

To add them, click on

the 'Darks' tab and add
the darks by clicking on:

Pixinsight will thus recognize

automatically the pre-
you too, Dark Flats thanks to the
indicated time "0.00s".
 The LIGHTs
We are going to add the lights
by clicking on the tab
Lights
Images need to be added to
process by clicking on "Add"
Lights.

ATTENTION: If the photos

were in colors, I would have
you select the matrix of
Bayer correspondent.

Let's see in each sub

for what needs to be chosen...

"Cosmetic Correction": on
car "Apply". As there is no
nothing in the drop-down menu
plant, it is necessary to close completely
closes the window by clicking
on "Exit". Don't worry,
everything remains automatically
recorded...

We are going to open the process/

Cosmetic Correctionand we are going
only check "Hot Sig-
to eliminate the pixels
hot lights.

We slide the icon (in

triangle) out of the window
Cosmetic Correction On
obtains a process called
Process01.
Now, we return
in the Script /Weighted
Batch Preprocessing

We select in the menu

unfolding our process...

In the section "DeBayer",

she remains inactive because it is
monochrome photos.

In the section 'Subframe

Weighing, I select
Generate subframe
weights. This allows the me-
sure quality of the photos
of the roundness of the stars...

In the sub-parameters,
I indicate my type of object.
Here, I choose 'Nebula'.

In the section "Image Re-

selection
"Apply", what will allow
an optimal alignment of
photos...
For the side "Image In-
tegraon", I select
Apply.

In the sub-section, I in-

dam for 'Rejection algo-
rithm: Linear Fit Clipping.

We have filled all the

settings.

Before launching the treaty-

we click on "Dia-
gnosics" to verify if everything
is OK...

We validate and we leave Pixin-

sight work...

Then, we open the directory

source of the files...

We open the directory "mas-

"and we click on the file
masterLight

There are 4 dark images that

open. It is enough to preserve-
see cell with the name "inte-
gra on. This is this photo
which will be addressed at the level
cosmetic...
 Subject: Pickering Triangle NGC 6979
Hubble processing and colors

CAMERA: ATIK HORIZON 2 40 x 180s

FILTER: IDAS NBZ (filter not suitable for this target)

WITH SELF-GUIDANCE
 CASE STUDY: NEBULA TREATMENT + HUBBLE COLORS

In this example, we
let's quickly deal with a
nebula captured by a ca-
my colors. Then, we
we change the color to
"Hubble colors" version.

We open the photo coming out of the

pre-processing. It is very
Sombre. It's normal.

I am going to click on this

icon to make appear
the content of the photo in
using the ProcessScreen
Transfer Function

I am requesting a green image (or

another dominant color.

Initially, we
let's crop the photo with
the processDynamicCrop

It is enough to draw the new one

rectangle on the photo in order to
cropper l’image.

To apply the crop, it

just click on:
Now, I am going to return
the more neutral sky background,
that is to say the closest
in a neutral gray. For this,
we equalize automatically
the color curves.

We use the process / Color-

CalibrationBackground-
Neutralization

We leave all the parameters

by default.

We get a black image.

It is enough to click on this one

icon to find a
quite a natural color of the
photo:
In my photo, I see very
good a gradient on the background
from the sky. This is explained by the
so the photo was taken
under the full moon...

To eliminate as much as possible

the gradient, in this case
I use the process/Dynamic
Background Extraction

I will click in the center of the

photo and I indicate these para-
meters
«Tolerance» : 1
«Samples per row» : 10

Then I click on 'Generate'.

I have small squares on

the photo, spaced in a way
fairly regular.

For each square piece,

I will click on it and
check that there is not a
star inside. If that is the
well, we need to move this square to
a place in the sky where there is none
no stars, because it distorts
withdrawal process of the
gradient.
If necessary, we can add
make squares by clicking
in dark areas of
the photo. And on the contrary, we
remove placed squares
on the nebula.

Then we indicate the para-

meters on the side.

We click on:

We get a black image

that we close, it's the gra-
It is said again.

But we keep our new-

want a photo with a gradient
strongly attenuated.
After returning the background of
neutral sky, we are going to cali-
to erase the colors of the photo.
For this, we use the process
Color CalibrationColorCa-
libraon

For the calibration of

colors are realized, it is necessary to
indicate 2 reference images
reference to Pixinsight.

White reference :
"Reference image" is a
"preview" of a clear area
so the nebula in uti-
saint the icon:

Background Reference :
"Reference image" is
a "preview" of an area
shadow from the bottom of the sky in
using the icon:

I also check "Struc-

true Detection

In the image opposite, we say-

Well, follow the 2 "previews":
Preview 1: Bright area of
the nebula
Previex 2: Dark Zone
from the depths of the sky.

We are starting the process in CLI-

how on the icon:
In the image on the right, we
you see the interest of the process
Color Calibrationto obtain
a better balance of
colors.

The photo shows a slight

green dominance. For the
delete, we use the Pro-
cessSNCR

We keep the parameters by

default and we apply it on
our photo by clicking on:
We are going to soften the photo
by reducing noise. For
this we use the process / Co-
lorCalibration /ACDNR

One must not overreact.

the sliders. They need to be positioned
give low enough notably
those opposite.

One must always protect well

the clear areas (nebula
and especially the stars) with the
parameter "Threshold".

We start the process in the CLI.

when on the icon:

One obtains a clearer image

sweet...
We are going to put me in ev...
dense in the nebula for
Adjust the sliders. To avoid
to saturate the stars, we
let's create a mask
stars using the Pro-
cessStarnet

This process does not work

in non-linear mode
so we must convert our
photo en mode non linéaire...

To do this, we open 2 pro-

cessHistogram Transformation
onandScreenTransferFunc-
on

We are going to slide the "triangle"

at the bottom of the window of
process HistogramTransfor-
maon.

The diagram no longer exists

linear.

In the Process /Histo-

gramTransforma on , on
apply the nonlinear in
clicking on:

The image becomes all

white. In the process
ScreenTransferFunction, on
click on:
Now that our image
is stretched, we can
create the star mask
with the ProcessStarNet

We are starting the process in CLI-

about the

We are going to get 2 images,

one with the nebula alone,
the other with the stars alone.

Let's set aside for a moment the

mask of stars...

In this case, I will

darken the sky background with
the Process /HistogramTrans-
formation

It is enough to move the cur-

seals under the graph...
In my case, I wish
obtain a photo of type
"Hubble", that is to say with
a slightly coppery red...

To do this, we will separate

make the photo in 3 colors.
We are going to use the Process /
Channel Extraction

We do not touch the para-

meters and we launch the pro-
cess.

I therefore have 3 photos in

corresponding grayscale level
in 3 colors: red, green and
blue.

I close the corresponding photo

gives to the blue.

I am going to create a new one

blue color from
red and green.

For this, I will use the

ProcessPixelMath

In the 'Expression Editor', I

you will write that we want the
blue color is a me-
long of 60% red and
40% green, that is the formula:
0.6*R+0.4*V

I check the box "Create new"

image» et je peux lui donner
a name...

I am starting the process and I ob-

Put on my "blue" blanket.
We therefore obtain our 3
colors.

Now, we will do
the inverse, that is to say combi-
mix the colors to create
a color photo.

We will use the Process /

Histogram Transformation

I indicate the information:

red layer
red layer
Blue layer
green layer

I do not forget to tick the

case "Chrominance Noise
Reduction.

I am launching the process in CLI.

how about:
I am therefore posting my photo
in colors with this red
coppery.

To adjust the photo, cut-

their brightest, background of
darker sky, I use the
ProcessHistogramTrans
formationby playing on the
cursors.

I will saturate a little the

colors to see better the
copper and blue... with the
ProcessColor Saturation

To see the result in time

real, we click on:

We are going to set up the line

horizontal, with the mouse,
to slightly saturate
the colors...
It is entirely possible
to mainly increase
the blue color by clicking
on the horizontal line at
level of blue. Then, it is necessary to
raise the curve...

There you go, he is still there

possible to work on the
nebula...

Now, we are going to

reopen the window with our
mask of stars...

Before adding our stars to

our photo, let's take advantage for
make a slight reduction.
For this, we use the Pro-
cessMorphologicalTrans-
formation

«Operator» : Erosion (mini-

mum

We launch the process in CLI.

how about:
Here are our stars.
small and thin.

there is nothing left to regret

the stars in our photo
with the ProcessPixelMath

The formula is simple

since we are accustomed to our
image + the star mask.