[gnuastro-commits] master 610986aa 08/23: Book: subsection of rgb-asinh script added to the color images tutorial

[Mohammad Akhlaghi]

branch: master
commit 610986aa61d50d3a4acb0a98b469fee6825f11e8
Author: Raul Infante-Sainz <infantesainz@gmail.com>
Commit: Mohammad Akhlaghi <mohammad@akhlaghi.org>

    Book: subsection of rgb-asinh script added to the color images tutorial
    
    Until this commit, the subsection of the tutorial for creating color images
    using the astscript-rgb-asinh script was missing.
    
    With this commit, this new section has been added. Right now it is at the
    first stage. It contains information that can be improved in terms of
    clarity and shortness. This will be done soon in next commits. In addition
    to this, the tips that are shown on the command-line has been changed in
    order to be more clear.
---
 bin/script/rgb-asinh.sh |  25 +++----
 doc/gnuastro.texi       | 195 ++++++++++++++++++++++++++++++++++++++++++++----
 2 files changed, 193 insertions(+), 27 deletions(-)

diff --git a/bin/script/rgb-asinh.sh b/bin/script/rgb-asinh.sh
index 1f3fb915..47c63fca 100644
--- a/bin/script/rgb-asinh.sh
+++ b/bin/script/rgb-asinh.sh
@@ -1052,22 +1052,19 @@ if [ ! x$quiet = x"--quiet" ]; then
   cat <<EOF
 
 TIPS:
-  # Use '--checkparams'to check the pixel value distributions.
-  # First use the default options to guess the parameters.
-  # Select a good sky background value of the images.
-      A minimum value of zero could be a good option: --minimum=0.0
-  # Focus on the bright regions and tweak '--stretch' and '-qbright'.
-      Try low values of '--qbright' to show the bright parts.
+  # First, use the default options to estimate the parameters.
+  # Select a good background value of the images:
+      A minimum value of zero could be a good option: '--minimum=0.0'
+  # Focus on the bright regions and tweak '--qbright' and '--stretch':
+      First, try low values of '--qbright' to show the bright parts.
       Then, adjust '--stretch' to show the fainter regions around bright parts.
       Overall, play with these two parameters to show the color regions 
appropriately.
-  # (next tips only for gray background image: --grayback)
-  # Change '--colorval' to select the value that separates the color and black 
regions.
-      --colorval->100 --> all becoming color.
-      --colorval->0   --> all becoming black.
-  # Change '--grayval' to select the value that separates the black and white 
regions
-      (from $grayval_guessed to 100.0)
-      --grayval->100 --> all becoming black.
-      --grayval->0   --> all becoming white.
+  # [next tips only for gray background image: --grayback]
+  # Change '--colorval' to separate the color and black regions:
+      Increase/decrease it to increase/decrease the color area (brightest 
pixels).
+  # Change '--grayval' to separate the black and gray regions:
+      Decrease it to increase the regions that are shown in black.
+  # Use '--checkparams'to check the pixel value distributions.
 
 PARAMETERS:
   Estimated: --qbright=$qbright_guessed --stretch=$stretch_guessed 
--colorval=$colorval_guessed --grayval=$grayval_guessed
diff --git a/doc/gnuastro.texi b/doc/gnuastro.texi
index 8e8f9f70..8c4af31e 100644
--- a/doc/gnuastro.texi
+++ b/doc/gnuastro.texi
@@ -8888,7 +8888,7 @@ Remember that there are a lot of pixels close to zero and 
only a few that are ve
 As a consequence, let's consider cutting brighter pixels.
 In general, a multiple of the standard deviation can be used for estimating 
this parameters.
 But we note the reader that this is very sensitive to each dataset, so be 
careful with the selection.
-Here, we will use @command{--fluxhigh 3.0} because it is roughly 3 times the 
standard deviation value of the i-band image.
+Here, we will use @option{--fluxhigh 3.0} because it is roughly 3 times the 
standard deviation value of the i-band image.
 
 @example
 $ astconvertt i.fits r.fits g.fits -g1 \
@@ -8899,8 +8899,8 @@ $ astconvertt i.fits r.fits g.fits -g1 \
 Open the color image and see that now the image shows some regions of the M51 
group.
 The most central part of the bright objects are visible.
 But still the vast majority of the image is in black.
-Play with different values of @command{--fluxhigh} to improve the image and 
obtain a better result.
-For example, with @command{--fluxlow 0.0} and @option{--fluxhigh 0.1} it is 
possible to obtain a color image in which the faint regions (pixels between 0.0 
- 0.1) are visible, but consequently, the bright parts are "saturated".
+Play with different values of @option{--fluxhigh} to improve the image and 
obtain a better result.
+For example, with @option{--fluxlow 0.0} and @option{--fluxhigh 0.1} it is 
possible to obtain a color image in which the faint regions (pixels between 0.0 
- 0.1) are visible, but consequently, the bright parts are "saturated".
 
 In some situations, depending on the scientific goal, this kind of images 
could be good enough.
 However, it may happen that other cases require even better color images in 
which the entire dynamical range is properly seeing, for example to visualize 
low surface brightness structures.
@@ -8936,22 +8936,191 @@ We have implemented a check in Warp to inform you when 
the images are not aligne
 
 @node Color image using asinh transformation,  , Color image using linear 
transformation, Creating color images
 @subsection Color image using asinh transformation
-Once the three channel images are at the same pixel grid, it is also possible 
to create a color image with the Gnuastro's @command{astscript-rgb-asinh}.
-The difference with respect to @command{astconvertt} is the transformations of 
the input images that is done in advance.
-Basically, the input images are transformed (by the asinh operator) in order 
to be able to show the entire range of pixel values.
-By running the following command, a color image with a gray background is 
generated,
+In the previous sections  we have aligned three SDSS images of M51 group 
@ref{Color channels in same pixel grid}, and create color images using 
@command{astconvertt}, @ref{Color image using linear transformation}.
+In this section we will show how to use the @command{astscript-rgb-asinh} 
script to generate color images.
+This scripts performs a non linear transformation of the input images before 
combining them to generate the color image.
+Basically, the input images are stretched using the asinh function in order to 
modify and show the entire range of pixel values as outlined by Lupton et al. 
(2004, @url{https://arxiv.org/abs/astro-ph/0312483}).
+See @ref{RGB asinh image} of this manual and Infante-Sainz et al. (2023, 
@url{TBD}) for more information.
+
+This script offers multiple options that you need to fine tune with the goal 
of obtaining the best color image quality possible.
+We will review in what follow some of the most important ones.
+Recall that images should be provided from redder to bluer, so in our case 
Red-Green-Blue means i, r, and g .fits images.
+Let's run the script with the default options on the SDSS M51 aligned images 
by executing the following command:
+
+@example
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --output m51-default.pdf
+@end example
+
+The script prints on the command-line some useful tips and the values that 
have been automatically estimated and used.
+We will follow and explain these tips in order to improve the output.
+Open the image and have a look, it is a color image with the background in 
black.
+Contrary on what happen by using simply @command{astconvertt}, now the images 
have been modified and you can be seen the M51 group.
+This output can be improved a lot!
+
+The first important parameter that we should properly set is the background 
value, or the minimum value to be displayed: @option{--minimum}.
+For each image it could be a diferent value, in that case you should provide 
comma-separated values with the option @option{--minimums}.
+In this particular case, a minimum value of zero for all images is good: 
@option{--minimums=0,0,0} or @option{--minimum=0}.
+
+@example
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 --output m51-min0.pdf
+@end example
+
+The difference with respect to the default image is not too much since the 
input images are quiete homogeneus, i.e., the background are zero and the noise 
around this value is small and similar in all of them.
+The default image appears more gray in the background, since it is trying to 
show the entire range of values.
+On the other hand, in the image generated with the minimum value of zero, the 
background appears more darker (strong black), since it is avoiding negative 
pixels to be shown.
+
+The next parameters that we should look at are: @option{--qbright} and 
@option{--stretch}.
+They control the asinh transformation that changes the pixel value 
distributions.
+The values that have been estimated is shown at the end of the execution of 
the script.
+Let's decrease @option{--qbright} by an order of magnitude in order to better 
show the very bright regions of the image.
+
+@example
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 \
+                      --qbright 1.481888e-02 \
+                      --output m51-min0-qbright.pdf
+@end example
+
+Open the image and check if the bright regions are properly shown.
+Now, decrease the paramter @option{--stretch} in order to show the regions 
around very bright pixels in linear scale.
+This allows you to show the fainter regions like the outer parts of the 
galaxies, spiral arms, stellar streams, and etc. as desired.
+
+@example
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 \
+                      --qbright 1.481888e-02 \
+                      --stretch 1.481888e-04 \
+                      --output m51-min0-qbright-stretch.pdf
+@end example
+
+Have a look at the image and check that now the faint regions are clearly 
visible.
+As you may have noticed, the ratio between the estimated paramteters are such 
as qbright/stretch = 10.
+This ratio is an emphirical value we have found after many color images 
generated.
+In the last example we set this ratio to be larger: qbright/stretch = 100, and 
it seems that it shows better the faint regions.
+There are other options that will improve this image but we will see them 
later.
+Now, let's use the option @option{--grayback} to create gray background images.
+In order to have a shorter acommand-line examples, we will use the internally 
estimated values for @option{--qbright} and @option{--stretch} paramters.
+
+@example
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 \
+                      --grayback \
+                      --output m51-min0-gray.pdf
+@end example
+
+Open the image and note that now the background is shown in gray!
+This color scheme is especially interesting for visualizing low surface 
brightness features.
+In this case, the very bright regions are shown in color, intermediate and 
faint regions are shown in black, and background or noisy pixels are shown in 
gray.
+Do you see the complex, diffuse and faint structures that are generated from 
the interaction of the galaxies?
+This was completely hidden in the linear or black background images, but now, 
just showing the background in gray with the option @option{--colorval} it is 
visible!
+The paramters that defines the separation between the color and black regions 
is @option{--grayval}.
+There is also another similar option that separates the black and gray 
regions, @option{--colorval}.
+
+First, try decreasing @option{--colorval} to 50.0 (by default it is 99.5) in 
order consider less area to be shown in color (only the very bright regions).
+
+@example
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 \
+                      --grayback \
+                      --colorval 50.0 \
+                      --output m51-min0-gray-colorval.pdf
+@end example
+
+For this value of @option{--colorval}, the estimated @option{--grayval} value 
is 96.2.
+Now, decrease this parameter to 70.0 in order to decrease the area that is 
shown in gray, or alternatively, increase the regions that are shown in black.
+
+@example
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 \
+                      --grayback \
+                      --colorval 50.0 \
+                      --grayval 70.0 \
+                      --output m51-min0-gray-colorval-grayval.pdf
+@end example
+
+By playing with this two parameters, it is possible to obtain the best result.
+The values used here are somehow extreme to ilustrate the logic of the 
procedure, but we encourage you to play with values close to the estimated by 
default.
+With the option @option{--checkparams} the script will show three histrograms 
on the command-line.
+They correspond to the different pixel value distribution that are used for 
estimating the parameters discussed until here.
+
+Instead of using a combination of the three input images for the gray 
background, it is possible to add a fourth image that will be used for 
generating the gray background.
+It is called the "K" channel and sometimes it is useful when a given filter is 
deeper or has some especial features.
+In this case the reasoning is the same as explained above.
+There are two other options that allow the smoothing of the different regions 
by convolving with a gaussian kernel.
+These two options are @option{--colorkernelfwhm} for smoothing the color 
regions, and @option{--graykernelfwhm} for convolving the gray regions.
+The value read by this options is the full width at half maximum of the 
gaussian kernel.
+
+Until now, we assumed that the three images have the same zero point magnitude.
+Said in other words, we considered that the pixel units were the same for the 
different channels.
+However, it may happen that the different channels have different zero point 
values.
+To account for this situations, the zero point values can be provided as 
comma-separated values with the option @option{--zeropoints}.
+In this case, the zero point magnitudes will be used for transform the 
different channels to the same units internally.
+
+In order to modify the color of the output image, it is possible to weight the 
three channels differently.
+This can be done with the option @option{--weights}.
+For example, with @option{--weights 1,1,4}, we are giving 4 times more weight 
to the blue channel than to the red and green channels.
 
 @example
-$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 --minimum 0.0 \
-                      --grayback --contrast 3 --output m51-asinh.pdf
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 \
+                      --grayback \
+                      --weights 1,1,4 \
+                      --output m51-min0-gray-bluer.pdf
 @end example
 
-In this example, zero has been set as the minimum value to be shown, we asked 
for a gray background, and we increased the contrast by 3.
-Have a look and check the differences with respect to the color image obtained 
above.
-Note that now it is possible to visualize the complete range of pixel values 
(from the brightest to the sky background).
-For more information about the @command{astscript-rgb-asinh} script and its 
options see @ref{RGB asinh image}.
+As a consequence, the output color image looks much bluer.
+This opens the possibility of alter the color of the images by a lot, and 
later analyses will be wrong.
+If the reduction, the photometric calibration, and the images represent what 
we consider are the red, green, and blue channels, then the output color image 
should be fine.
+In some situations it may happen that the combination of channels has not a 
traditional color interpretation.
+For example, the result of combining an x-ray channel with an optical filter 
and a far infrarred image has a complicate interpretation from the point of 
view of what humans understand by color.
+However, the physical interpretation is that the different channels (colors on 
the output) provides different physical phenomena of the astronomical sources.
+Overall, we recall that this option should be used with care.
+With great power there must also come great responsibility!
+
+There are two other transformations that allows to change the output color 
image appearance.
+A linear transformation that consists in a combination of adding brightness 
and increasing the contrast with the options @option{--brightness} and 
@option{--contrast}.
+Usually, only the contrast is enough to improve the quality of the color image.
+For example, generate the previous color image and compare it with the image 
using @option{--contrast 3}:
 
+@example
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 \
+                      --grayback \
+                      --output m51-min0-gray-default.pdf
+
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 \
+                      --grayback \
+                      --contrast 3 \
+                      --output m51-min0-gray-default.pdf
+@end example
+
+As it is clear, the contrast and clarity of the image with the higher contrast 
is much better.
+A gamma correction (non linear) is also available through the option 
@option{--gamma}.
+This is a different type of transformation that could be useful depending on 
the particular case.
+Try different values of gamma and check the results with respect the default 
values image.
+Low values of gamma will show fainter structures on the image, while higher 
values will show brighter regions.
+
+@example
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 \
+                      --grayback \
+                      --gamma 0.25 \
+                      --output m51-min0-gray-gamalow.pdf
+
+$ astscript-rgb-asinh i.fits r.fits g.fits --hdu 1 \
+                      --minimum 0.0 \
+                      --grayback \
+                      --gamma 2.2 \
+                      --output m51-min0-gray-gamahigh.pdf
+@end example
 
+Through this tutorial, we have tried to reproduce the basic steps in order to 
construct a color image from three different images using the 
@command{astscript-rgb-asinh} script.
+We note that the parameters that generate the best color image will depend a 
lot on what you want to show, and the quality of the input images.
+We encourage you to follow the tutorial step by step with the SDSS images, and 
later with your own data.
+See @ref{RGB asinh image} for more information, and please cite Infante-Sainz 
et al. (2023, @url{TBD}) if you use this script.
 
 @node Zero point of an image, Pointing pattern design, Creating color images, 
Tutorials
 @section Zero point of an image