The postprocessing 2 - Color Calibration Our goal is simple yet ambitious: we want to create an image of the planetary nebula in color, which is as close as possible to view "real" object, just as if we were floating in space near of M27.
using any image editing program, we soon realize that the manipolazione dei colori così come della luminosità o del contrasto, ci porta a creare centinaia di modalità di visualizzazioni, a volte molto differenti l'una dall'altra.
Questi potenti strumenti informatici, più che aiutarci, così come avviene nelle fotografie terrestri tradizionali, in questo caso ci disorientano e ci lasciano perplessi: quali sono i veri colori degli oggetti celesti?
Le nostre incertezze sono dovute al fatto che non abbiamo nella nostra memoria visiva un'immagine "reale" di confronto, come può avvenire per gli oggetti terrestri: il
rosso "Ferrari" , per fare un esempio, sappiamo com'è e possiamo più o meno avvicinarci ad esso se abbiamo necessità di equilibrare i colori di una foto sportiva fatta ad un Gran Premio. Con tecniche più oggettive, i fotografi professionisti utilizzano la tecnica del "bilanciamento del bianco" che consiste nell'equilibrare i tre canali RGB fondamentali in modo tale che un oggetto bianco o grigio abbia gli stessi valori di pixel nel rosso, nel verde e nel blu.
Ma il rosso delle emissioni
H-alfa presente in tante nebulose (compresa questa) come appare nella realtà? Anche con telescopi di grande apertura e sotto cieli incontaminati è difficilissimo cogliere i reali colori di nebulose galassie e comete perchè i nostri occhi lavorano in regime di bassissima illuminazione: siamo in una situazione di
visione scotopica : i nostri eyes have difficulty distinguishing differences in chromaticity and everything seems almost "black and white" (and this can be also found in terrestrial normal night vision).
But just as there is a reference to "ground" for white, even in astronomy, we can use a "light blue" reference: our Sun, or rather, the stars, like our Sun, are of a very precise and determined the spectral class
: the
G2V . Just like the sunlight at noon (well, at the zenith) is considered the standard for white light earth, the stars of the same spectral class G2V may be taken as a reference standard for white in astronomical images.
A list of these reference stars you can find it here
, along with the file. Tdf to identify them in the program
GUIDE , truly global multi-functional and cost often use to support my observations.
Our set is images of M27 however, was not acquired with the intention of making use of the tri-color "aesthetic," so we have not taken one of these stars using the same instrumental set, we must be content with what we (often the case, unfortunately!)
To continue this short tutorial, but we can make a reasonable approximation: a star search in the field of view spectral type close to that reference.
Let's help once again GUIDE: This program can represent the stars with the color corresponding to the spectral class of membership (see above) and simply click the right button on the star to read the information available: the Class G stars are represented in yellow, and we immediately note that there may be two candidates to the east of the nebula. That lower, very bright, we must immediately discard as it is always overexposed, while a bit 'higher up, just below the variable DQ Vulpeculae, is nearly perfect: it is the HD 345452 and class spettrale G5, un pizzico più "arancione" del nostro Sole. Certamente questa piccola differenza non avrà un effetto tangibile nel nostro processo di calibrazione colore.
Abbiamo un'ulteriore fortuna: al momento della ripresa del set di immagini, M27 era ad oltre 70 gradi sopra all'orizzonte, quindi risulta praticamente nullo ogni effetto di arrossamento dovuto all'assorbimento atmosferico.
Ora che abbiamo tre immagini RGB equivalenti (stesso tempo d'esposizione) dello stesso oggetto ripreso con lo stesso identico set strumentale e per di più nella stessa serata, senza che l'oggetto abbia cambiato significativamente la propria altezza sul cielo, possiamo procedere al calcolo dei coefficients for color correction with the reference star HD 345452. The operation is very simple: for each image R, G and B flows are measured in ADU of the star sample, for example using the zoom window and using the pointer to "crown".
In the above we can see is the measurement of the stars: you choose a range for the pixels in central opening so as to contain most of the stellar flux (usually chooses an aperture equal to 4-5 times the FWHM
of the evening). The status bar of the window are three numbers V, S and B: these are: the net flow of the full sky, flow includes the sky background and finally the local value of the background sky. Note the value V for the three images in R, G and B:
VR = 334173 = 211426 VG
ADU ADU ADU
VB = 103364
calculate the weights of the colors than, for example, the red channel we get
R = 334173/334173 = 1.00
G = 334173/211426 = 1.58
B = 334173/103364 = 3.23
Substituting these three coefficients corresponding to those calculated automatically by the dialog
Astroart Tri , finally we get the picture with the colors balanced with respect to a star of spectral type similar to our Sun (see below).
If we compare the image above with the balanced automatically by Astroart will notice a difference: in this case, the automatic balancing favored the green and blue channels.
image above, properly balanced with a star of spectral type similar to that of the Sun, most stars appear yellow-orange that used for the calculation of the coefficients appears perfectly white, while investigating further, there are some stars with very special colors, from bright red to blue to blue (see image below).
1 - Variable
DQ Vul: it is a giant red, variable type Mira
. The white star in the low to his right is our reference star HD 345452, which is used for color balance.
2 -
central star of M27
is a white dwarf and is a decided color azzurognolo.
3 - Variable "
Goldilocks" is probably a long-period variable of the class "Mira". The feature that distinguishes this star is the curious circumstance that led to his discovery: do you think has been identified for the first time by an amateur astronomer in the Czech Republic comparing ...
two magazine covers astronomical !
4 - A mysterious star that is characterized by the strong blue color.
In the chart above, a collection of spectra of stars of different spectral class: each spectrum has been separated in the scale of the relative brightness to avoid duplication (data taken from Pickles 1998).
As you can see the stars emit light across the visible spectrum band: the color (ie surface temperature) is determined by the position of peak brightness.
planetary nebulae, however, as well as supernova remnants or areas of star formation in galaxies, emit radiation bands much more isolated and restricted: it is said that
have a range of "issue" characterized by rows of bright
generated by ionized elements, especially hydrogen and oxygen. The color of such objects is therefore strongly influenced by these "gaps" in the emission of radiation and especially from our "vision system" or by the telescope, the CCD camera, and especially by the filters used for recovery.
To realize this mechanism is more convenient to construct a chart: 100 made a hypothetical emission intensity of the spectral line, or transmission of a filter or quantum efficiency of our CCD camera, we can draw respectively
lines most important issue of our planetary nebula (Hbeta to 486.1 nm, OIII to 495.9 and 500.7 nm and 656.3 nm in Halfa) represented by the continuous vertical lines;
the transmission curves of the filters B, V and R represented by the dotted lines;
of CCD quantum efficiency (QE) represented by a continuous curve of fuchsia.
The emission line two times ionized oxygen on OIII 500.7 nm is by far the most intense and therefore the one that dominates the brightness and color of the nebula. The second most important line is the intensity as the transition to hydrogen H alpha 656.3 nm. We note that the latter falls almost exclusively in the area of \u200b\u200bthe red filter, as it should happen as it is at a wavelength of the light spectrum decidedly in the red.
different and more delicate is the situation for the row of 'OIII about 500.7 nm: note that falls in the area of \u200b\u200binterference of the two filters B and V: in fact, there is also shown in color scale in the x-axis, 500 nm almost exactly the area seem to separate from the blue green color spectrum. Unfortunately, this dichotomy is not perfect just as well represented by the filters used: in fact the line of 'OIII filter V is intercepted by around 50% of its relative intensity. The same line but was intercepted by B filter only 7% of its intensity on! This is why the nebula appears to us so green!
To bring a little 'things we must give the same weight of the green to blue filter, ie, a simple ratio B = 0.5 / 0.07 = 7.14.
Back then 3 images in our RV and B of the nebula and with the command
Tricromia Astroart apply to the following factors:
R = 1.00 G = 1.00 B = 7.14
The result? Here it is below.
If the two images side by side centered on the nebula, we can better appreciate the differences.
So, in summary, we have attempted an initial color calibration using the technique of spectral type star sample similar to our Sun which is a G2V star. The stars appear with colors consistent with their spectral classes but maintained a strong nebulous light green: so in real life? Probably not. In fact we have seen that these nebulae, unlike stars, they emit primarily on well-defined lines of the electromagnetic spectrum. In particular, M27 has a strong emission line three times ionized oxygen [OIII] which has the distinction of emitting around 500 nm, ie in the middle of the transition from blue to green spectrum.
Everything depends on how they act out the filters B and V (= G) used .
In particular, the V filter used for this set of shooting, trapping at least 50% of the issue, unlike the B filter that records only 7% (for the same course of time of exposure). To balance the colors (ie for 50% the contribution of the blue filter), we then multiplied by the coefficient of the color blue filter for an appropriate correction factor (in our case B = 0.5 / 0.07 = 7.14 while keeping R = 1 and V = 1). As
each filter is different (even the same filters of the same manufacturer are slight differences) should be compared to determine the influence of these OIE and redefine the line ratios of the RGB color.
so doing, however, we return to unbalance the colors of the stars.
It follows that, with only three RGB images, we could never make an image with stars and nebula simultaneously balanced colors. Actually, you may correct the image of the heavy editing (though this is not our goal), or, even more complex and expensive, can be corrected using a fourth image, made with a narrow-band filter on the 500, 7 nm, or to be added in the band in blue or green band, after the color calibration of the method G2V star. We
Anyway, here on a composition that should give the idea of a composite image of coins. 
It is clear that the values \u200b\u200bare more prevalent than those of 1 and 2 cents, while less frequent by 5 cents and 2 euros. 
The first picture is a classic popular astronomical object: a comet. Even in the case of other entities such as galaxies or nebulae, however, the histogram does not differ much from that mostrato sulla destra. Abbiamo una forte concentrazione di pixel poco luminosi, ovvero sulla sinistra dell'istogramma seguito da un graduale presenza di pixel via via più luminosi. La prima concentrazione altro non rappresenta che il numero di pixel del fondo cielo: non a caso il picco si attesta attorno ai valori 12-13 ADU, proprio quelli indicati come valore di fondo cielo dell'immagine (vedi il valore B nella barra di stato sotto l'immagine). La seconda parte dell'istogramma rappresentato dalla "coda" di valori via via più luminosi identifica la chioma e la coda della cometa con le stelle comprese nel campo di ripresa. Si noti poi il picco intorno al valore 255 ovvero il valore più luminoso per un'immagine a 8 bit: si tratta di tutti quei pixel concentrati in the more luminous (and often saturated) image as the center of the comet's coma and the saturated star in the upper right. 
Finally, as a third example, a typical figure ground: a human face. In a correctly exposed photograph and includes a balanced view of the entire range of gray values \u200b\u200bfrom 0 to 255 with the appearance of the histogram is almost flat. 
