Examining the morphology of the Ring Nebula

(M57) through CCD photography

João Quintans, Ricardo Branco, Rodrigo Salvado (ist1106267, ist1106162,

ist1106057)

MEFT, Instituto Superior Técnico

University of Lisbon

2025
 Abstract
This work aims to study the planetary nebula M57 (the Ring Nebula) through photometric and astrometric data analysis,
using CCD imaging and standard BVR filters. The main objectives were to obtain calibrated magnitudes, construct a
physically meaningful color-composite image, and characterize the nebula’s morphology through surface brightness analysis.
Observations were carried out with the Celestron Schmidt-Cassegrain telescope and a Moravian CCD camera at the CIGeoE
Observatory in Lisbon.
The raw images were processed through a Python-based data reduction pipeline including bias and flat-field correction,
image alignment, stacking, and background subtraction. Photometric calibration was performed using the nearby standard
star Sheliak, allowing the conversion of instrumental magnitudes into calibrated magnitudes. The resulting RGB image
revealed the typical morphology of M57, with a bright red ring dominated by ionized nitrogen and hydrogen and a green
inner zone of doubly ionized oxygen. However, it failed to accurately portray the ionized helium, responsible for the the blue
region in the centre.
The surface brightness profile showed the expected decrease in intensity toward the nebular cavity, with maximums at
the ring radius of about 70 pixels, and in the centre, where the white dwarf lies. The astrometric calibration, performed
using the [Link] service, yielded a final image scale of 0.481 arcsec/pixel and the nebular center coordinates
RA = 18h 53m 32.258s, Dec = +33◦ 02′ 09.53′′ . The central white dwarf was successfully detected and measured through
iterative PSF photometry, giving calibrated magnitudes of B = 15.405, V = 15.614, and R = 16.357, in close agreement
with SIMBAD values.

ized nitrogen. Inside this region lies a hotter zone of ionized

1 Introduction
oxygen, glowing green, while at the center, blue lobes of ion-
A planetary nebula is an astronomical object formed dur- ized helium extend perpendicular to the main ring. Along
ing the late evolutionary stages of low to intermediate mass the periphery of the nitrogen zone and the start of hydrogen
stars. During the main sequence phase, the longest in stellar rich area (visible in a lighter red), dense dark knots of gas
evolution, these stars burn hydrogen in their cores maintain- have persisted the stellar winds, casting shadows onto the
ing hydrostatic equilibrium. When the hydrogen in the core is surrounding hydrogen shell. [3]
exhausted, the star evolves into the subgiant branch and later
becomes a red giant. The star then undergoes a strong mass
loss, expelling its outer layers into space. Meanwhile, the stel-
lar core contracts and heats up, radiating intense energy that
ionizes the previously expelled gas, causing the atoms within
it to emit light. The resulting shell of glowing, ejected mate-
rial is what we observe and named as a planetary nebula. [1]

1.1 Historical context and composition

Despite the name, planetary nebulae have no relation to
planets. The term was introduced in the 18th century, when Figure 1: Structure of the Ring Nebula (M57). (NASA) [3]

observing these objects with small telescopes, astronomer

William Herschel thought their blue-green glow resembled the
newly discovered planet of Uranus.
1.2 Motivation and scientific goals
Messier 57 (M57), more commonly known as the Ring Studying planetary nebulae like the M57 is particularly im-
Nebula, is located in the constellation Lyra at a distance of portant because they represent a fundamental ending phase in
roughly 2,000 light-years. It was discovered in 1779 by thethe life cycle of low to intermediate mass stars, which include
French astronomer Antoine Darquier de Pellepoix. With an our own Sun. Observations of its morphology and ioniza-
apparent magnitude of 8.8, the nebula is easily observable tion structure provide valuable insight into stellar mass loss
with small to medium-sized telescopes. In its center lies a processes, the dynamics of gas ejection (whose emission lines
white dwarf, the stellar remnant of the progenitor star, with reveal relevant physical parameters such as temperature and
an apparent magnitude of about 15.0. [2] density) and the enrichment of the interstellar medium with
Regarding its physical composition and structural appear- heavy elements, which will later become prime material for
ance, the Ring Nebula exhibits a red ring dominated by ion- other new stars and planets to be born. [4] [5]

2
 In this work, our first goal is to measure the flux of M57 of 10 parsecs from Earth [7]. In this work only the apparent
and produce a color-composite RGB image with physical sig- magnitude will be relevant to us, which is given by:
nificance. While achieving this, we will also determine cali-  
F1
brated magnitudes for this target with help of a calibration m1 − m2 = −2.5 log10 (1)
F2
star. In addition, we aim to obtain the final image with an
When we are calculating the effective magnitude, S, of a
astrometric solution, including the World Coordinate System
extended object in the sky, such as a nebula, this is, not a
(WCS). Furthermore, we will study the morphology of M57
point source, we need to make a correction to account for the
by analysing a surface brightness profile, measuring how the
area of the object:
flux intensity in different filters varies with distance from the
centre of the nebula’s elliptical structure. Finally, we will S = m + 2.5log(A), (2)
attempt to identify and measure the magnitude of the cen-
tral white dwarf of M57 using point-spread function (PSF) where A is the area of the nebula in arcsec2 .
photometry.
Since this nebula itself is relatively compact (1.4x1’), stan-
2 Observations and Data
dard CCD imaging with narrowband filters will allow suffi-
cient resolution to investigate the structural details and ob- The CIGeoE Observatory is located in Moscavide, Lisbon
tain a decent quality color-composite image, despite the chal- (38.7784° N, 9.1140° W). Although the observatory is affected
lenging light pollution conditions of the Lisbon metropolitan by the light pollution typical of the Lisbon metropolitan area,
area. Also, since this is a well studied stellar object, it makes and the nearby airport, the site is well suited for training in
it easier to find accurate reference measurements of its most astronomical observation and instrumentation.
relevant properties. In our observations we used the Celestron Schmidt-
Cassegrain telescope present in the CIGeoE observatory, as
1.3 Useful concepts illustrated in figure 3. The principal specifications of this
instrument are summarized in Table 1.
In astronomy, the position of a object on the celestial
sphere is usually expressed in the equatorial coordinate sys- Aperture Diameter Secondary Mirror Diameter Focal Lenght
35.5 cm 10.0 cm 391 cm
tem, which is basically the projection of the latitude and lon-
gitude coordinate system onto the celestial sphere. In this Table 1: Characteristics of the telescope used in the observation
system, lines of latitude correspond to lines of declination
(Dec, δ), measuring angular distance north or south of the The system is placed on an equatorial mount, which allows
celestial equator in degrees. Lines of longitude correspond to the tracking of celestial objects by compensating the Earth’s
right ascension (RA, α), measuring angular distance eastward rotation. The optical efficiency is approximately 100%, ensur-
along the celestial equator from the vernal equinox, usually ing that nearly all the incoming light is transmitted through
expressed in hours, minutes, and seconds (ranging from 0h to the system.
24h) [6]. Figure 2 shows a schematic representation for better
interpretation of the coordinate system.

Figure 2: Right Ascension and Declination [6]

The brightness of a star is measured in terms of the radiant Figure 3: CIGeoE’s Celestron Schmidt-Cassegrain telescope
flux F received. We can distinguish two types of magnitudes.
The apparent magnitude (m) quantifies how bright an object In order to perform the acquisition of imagery, we used
appears when observed from Earth. On the other hand, the a Moravian CCD camera. This instrument is optimized for
absolute magnitude (M ) is the apparent magnitude that an astronomical imaging, providing features particularly impor-
object would have if it were located at a standard distance tant for long-exposure observations of deep-sky objects such

3
as M57. The main technical specifications of the camera are Using the Astronomical SNR calculator [11] and inputting
summarized in Table 2. the parameters of the camera and telescope, it was possible
to determine the optimal exposure time that avoids CCD sat-
Parameter Value
uration while also maximizing the SNR. In this case, initially,
Size (pixels) 1536 × 1024
we had planned a high exposure time, but due to issues with
Pixel size 9 × 9 µm
the telescope’s tracking of Earth’s rotation, it had to be re-
Readout noise 16 e− rms
duced. We therefore settled on 10 frames with 30 seconds
Dark current 0.19778 pA/cm2 (at T=0◦ C)
exposure time for M57, as stacking these images, which is
Gain 1.5 e− /ADU
discussed in data analysis, will give us a good enough SNR
Quantum efficiency 70%
Tdoubling 6.3 ◦ C to work it. For Sheliak all exposure times will be less than a
Saturation threshold ∼ 55000 ADU second per frame.
Operating temperature ∼ -10 C
◦ Table 4 shows us some useful informations about the tar-
gets and estimated time per target.
Table 2: Main characteristics of the MORAVIAN camera [8]
Target α (J2000) δ (J2000) Time on target Mag
M57 18h 53m 35.16s 33◦ 01’ 43.20” 35 minutes m = 8.8 / S = 15.9
As discussed on section 1, we will focus primarily on M57. Sheliak 18h 50m 04.80s 33◦ 21’ 45.60” 5 minutes m = 3.55

During the night chosen for the observation, 18th of

Table 3: Observational data for M57 and calibration star, Sheliak. α
September 2025, M57 was near zenith, reaching its highest describes the right ascension and δ stands for the declination. S is the
altitude at 20:40 (altitude ≈ 84.5◦ , airmass ≈ 1.01) with no effective magnitude given by equation 2.
Moon visible above the horizon, as seen in figure 4.
When also accounting for the time required for telescope
slewing, pointing, focusing, and changing filters, as well as the
additional time needed to cool down the CCD and to track
our targets, we estimated a total of about 2 to 2.5 hours at
the observatory. However, due to cloudy conditions upon ar-
rival, we had to wait before observations could begin, which
extended our stay to approximately 4 hours. Table 4 summa-
rizes all the exposure times of the observation.

Image Type Object Filter N texp [s]

Bias – – 10 0
B 10 0.5
Figure 4: Daily chart for 18th September 2025 [9] V 10 1.5
Flat Sky
R 10 1.5
A calibration star is also required to obtain calibrated mag- Black 10 0.1
nitudes, which are essential for producing a final color image B 10 0.5
with physical significance. For this purpose, Sheliak was cho- V 10 0.2
Calibration Sheliak
sen because of its proximity in the sky to M57 and because R 10 0.1
there were available catalogued magnitudes for all the filters Black 10 0.1
used in SIMBAD database. [10]. B 10 30
For both M57 and Sheliak, a set of images is required with V 10 30
Science M57
no filter as well as with the three filters used: blue B, green R 10 30
V, and red R. In addition, Bias and Flats frames are also Black 10 30
necessary for data reduction, which will be discussed further
Table 4: Observation log for calibration and science frames. N repre-
in Section 3.1. sents the number of images.
The next step was calculating the necessary exposure time
for each filter. For this, we need to take into account that Finally, it is important to note that although the positions
when using a CCD camera we must not exceed, or even ap- of our targets in the night sky are known (as shown in Figures
proach, the saturation threshold, which is about 55,000 ADU. 5 and 6), poor visibility required us to use a bright reference
However, the signal-to-noise ratio (SNR) should be at least star together with the telescope’s tracking system to calibrate
greater than 25 to obtain a usable image for photometry, al- the pointing, allowing us to locate M57 and Sheliak more
though the resulting quality will still be relatively low. easily.

4
 Data reduction starts by stacking the Bias files, creating
a Master Bias. Similarly, stacking the flat frames yields a
Stacked Flat, to which the Master Bias is subtract, creating
the Master Flat, a flat composed of the medians of the 10
flats taken, corrected for the electronic noise [12]:

Master flat = Stacked flat - Master Bias (3)

The master flat can be seen in figure 7 (b). When looking
Figure 6: Position relative to Vega. closely, it’s possible to notice that figure (a) is more "grainy"
Figure 5: Sky View of M57 Sheliak is represented by β (β Lyr) than figure (b). This extra graininess is due to the bias noise.
In figure 8 it is possible to see the difference between both
images- the bias file. Subtracting this frame reduces the
3 Methodology graininess of the original stacked flat.

3.1 Data reduction

The astronomical data obtained from the stellar objects
observed must be treated before being analysed, in a process
called Data Reduction, where noise, unexpected fluctuations
and inefficiencies of individual pixels must be accounted for.
For this, two extra acquisitions are taken, whose only purpose
is to correct these corruptions: the bias and the flats [12].
The bias consists of an acquisition with no exposure time.
In theory, no readings are to be expected, but in practice,
Figure 8: Bias acquisition
electronic noise can produce counts in some pixels [12]. The
flats are acquisitions of surfaces of constant luminosity- usu- The Master flat now contains information about the re-
ally taken during twilight, when the sun has set, but the sponse of each pixel to the same light. As such, by normaliz-
sky is still illuminated- with some exposure time [12]. These ing the light of our stellar objects to the values of the flats, we
frames reveal the relative response of each pixel to the same can compensate for pixels with reduced activities. So to cal-
incoming light, thus exposing pixel inefficiencies, dust on the ibrate our Stacked Science images, we must first remove the
telescope optics, and other factors that prevent a uniform il- electrical noise, and then normalize each pixel to the value
lumination across the detector. Flats are obtained for each obtained in the flat [12]:
filter, as well as without any filter. In figure 7 (a) we can
see the flat taken with no filters. The spots visible are dust Stacked Science - Master Bias
= Reduced Science (4)
specks in the telescope mirrors. Master Flat
The Reduced Science image is closer to the actual object
than the raw data. By normalizing the value of each pixel
by the flat pixel value, we can compensate the pixels that
register less light and correct the final image.
In figure 9 we can compare the two images of the neb-
ula M57 before (a) and after (b) Data Reduction. We can see
that the spots of dust disappear, and that in general, the light
(a) Flat 15, no filter (b) Master Flat, no filter distribution of the background is more homogenous, and the
subtle round cloudiness that appears in the centre of the raw
Figure 7: Comparison between raw and master flats
file disappears on the reduced image. The nebula itself does
not change substantially, still remaining close to the raw im-
Ten acquisitions were taken for each file type—bias, flats,
age, so we can see that these corrections are not very extreme,
and the science images of the stellar objects of interest (She-
but still significant.
liak and M57). The value of each pixel is computed as the
median of its ten recorded values across the images, which
3.2 Data analysis
helps suppress random errors and transient fluctuations, such
as those caused by cosmic rays. This process is known as A long process of data analysis is done to get quality images
stacking. for analysis from the raw data. This process undoubtedly

5
 Figure 10: First (left) and tenth (right) acquisitions of M57, no filter,
side by side. White lines connect the same point in both images.

(a) Raw M57, no filter, 30 seconds exposure

When aligning the images, these are all "dragged" to the
same reference point, chosen from the first acquisition. When
a certain image is dragged, some part of it is cut out, and at
the same time it creates an area where there are no pixels.
This can be seen in figure 20 - there is a black edge on the left
and bottom of the image that resulted from the frame being
dragged upwards and to the right. This strip of pixels that
did not exist before is set so that every pixel has value zero,
hence the sharp contrast in color from the rest of the pixel,
with non-zero counts.
(b) Reduced M57, no filter, 30 seconds exposure

Figure 9: Comparison between raw and reduced M57 acquisitions

starts with data reduction, explained in section 3.1.

Figure 11: Stacked M57, no filter

3.2.1 Reduction, alignment and stacking
After aligning the files to the reference point, the images
are stacked, once again by taking the median of the 10 aligned
The first stage of data analysis consists of reducing, aligning
acquisitions for each pixel, and thus concluding the first stage
and then stacking files.
of data reduction. After this stage we obtain the stacked
As explained in section 2, since high exposition times were image of the object of interest. This process is repeated for
not achievable with the conditions and equipment available, acquisitions of different filters.
sets of 10 expositions of 30 seconds were taken for M57, and
similar for Sheliak, but using very short acquisition times. 3.2.2 Background removal

To treat this data, each raw file is first reduced. After The second step to treat our data effectively is to remove
this step, it is necessary to align the files. The system used to the background noise. For this purpose it is used photutils-
compensate for the earth’s rotation was not working properly, a python library that has a class called Background2D. This
so between the first and the tenth acquisition of a stellar class can estimate the background using statistics, and by
object, its position in sky would change, and in the CCD as fine-tuning initial parameters, we can obtain a good back-
well. This can be confirmed in figure 10. Side by side we can ground estimate, to later remove it.
see two acquisitions of the nebula M57- the first acquisition However, a subtle step must be accounted for before re-
taken on the left, and the tenth on the right. Though subtle, moving the background. The strip of pixels created after
there is a shift of the image to the lower left corner. The aligning the images all have value zero, and this could affect
white lines in the image connect the same points in the both the method used by the python function. To prevent this
images. If no shift happened, these lines would be straight, from ruining the results, it is used something called a mask.
but they are not. A threshold value is defined, and every pixel that has a value

6
lower than the threshold is assigned a mask. Pixels that have 3.2.3 Aperture photometry
a mask are not accounted for when the background estima-
Aperture photometry is an essential tool used by as-
tion is calculated, as to not disrupt this process. Since the
tronomers to analyse stars. With it, the flux of stars and
pixels from the strip all have value zero, a very low threshold
other stellar objects can be calculated, thus allowing the char-
can be chosen that targets only those pixels.
acterization of various properties of these objects.
In table 5 it is possible to observe the median background, We first need to define our aperture- the region of space
the median root mean square (rms) background estimated, for which we want to calculate the flux. A circle is chosen,
and the percentage of pixels masked for all filters used. The both for nebulas and stars, since it approximates their shape
filters that had the higher counts- red and green- also had quite well. By getting the coordinates of the centre of the
the highest median background, and the median background object of interest, and then drawing circles of varying radii,
of the three filters (548) sums to a value close to the median we can observe the flux as a function of the radius of the
value obtained without any filter (554). aperture. This can be seen in figure 13. On the top image, the
different circular apertures are drawn over the star Sheliak.
Filter
Median Median rms Fraction The bottom image shows the evolution of the flux with the
Background Background Masked (%)
radius. It is to be expected that the bigger the aperture,
No filter (Black) 554 6.29 2.2
Red 312 4.93 1.7 the higher the flux value, and in fact that is what happens.
Green 201 4.31 0.3 However, this value plateaus after a certain radius.
Blue 35 2.81 0.2

Table 5: Characteristics of background for each filter

On figure 12 four images are represented regarding the neb-

ula M57 taken with the red filter. From the two images on
the top it’s possible to compare the acquisition before and
after the removal of the background. The background esti-
mated can be seen on the bottom left, and its correspondent
error on the right. The background estimated has a cloud
like shape- it is more present in the middle of the image, and
it fades closer to the edges. It is important to note that the
background tool did not remove any of the stars present on
the image, which is a good indicator that the background
subtraction for the nebula points, our object of interest, was Figure 13: [Top] Apertures of varying radii plotted over Sheliak star,
taken with red filter. [Bottom] Flux inside circular aperture as a function
not overestimated, and the object remains "intact".
of aperture radius

After the whole star is within the aperture, increasing the

radius only adds background to the sum of the flux. But
since the background was removed, adding background to the
aperture does not increase flux, or at least it only increases
it by a residual amount. The radius of the aperture was
chosen accordingly- auxiliated by the bottom graph in figure
13, the radius was picked in the point where the flux stopped
increasing rapidly. In the case of figure 13, the chosen radius
was 14 pixels.
For M57, the exact same procedure is adopted, the only
difference being the radius chosen is much bigger, having gone
with 80 pixels.

Figure 12: [Top left] Stacked M57, red filter. [Top right] Stacked M57,
3.2.4 Color-composite (RGB) Image
removed background, red filter. [Bottom left] Estimated background,
red filter. [Bottom right] Error of background calculated.
After completing the reduction and image processing treat-
ments, the next objective was to produce a color image with

7
physical significance, meaning that the contribution of each However, we need to also take in consideration the extra
filter reflects its correct proportion of color. term (mZP ) in equation 7. After some simple algebra:
Firstly, we used aperture photometry for both M57 and
Sheliak to obtain the flux of the objects and then compute
FZP = 10−0.4mZP (11)
the instrument magnitude (in arbitrary units) which is given
by:
Concluding, the correcction we need to make to each pixel
in the image associated with the filter i is:
 
Finst
minst = −2.5 log , (5)
texp
where Finst is the flux of the object, obtained from the k[i] sec z
Fcalib [i] = 10 2.5 × F inst[i] × FZP [i] (12)
image, in arbitrary units (ADU).
If the target and standard (calibration) star are not in the
After this flux correction, we can finally compile the 3 im-
same field, they will have different airmasses, which means
ages (B, V and R filters) using the make_rgb function from
that atmospheric extinction is not the same. In this case we
Astropy library [13]. Figure 14 represents the two results.
will need to make a correction given by:

minst,0 = minst − k sec z, (6)

where sec z is the airmass at the time of observation [9]

and k is the extinction coefficient, a parameter that varies
with each filter. Table 6 shows the values of the extinction
coefficient for the the filters used.

Filter B V R
k 0.25 0.15 0.09

Table 6: Extinction coefficients for each filter

(a) M57 Color RGB - LogStretch(a=200)

Using minst,0 its possible to obtain the calibrated magni-

tude of the targets:

mcalib = minst,0 + mZP , (7)

where mZP is the magnitude of the zero point (ZP), which

tell us how much to shift our instrument magnitudes to match
the standard system. This is given by:

mZP = mstd − mstd,0 , (8)

where mstd is the catalogued magnitude of the star in the (b) M57 Color RGB - AsinhStretch(a=0.005)

corresponding filter, which can be obtained from the SIM- Figure 14: Comparison between the two RGB images obtained with a
BAD database [10] shown in table 7, and mstd,0 is given by different type of stretch
equation 6, applied to Sheliak.
Starting from equation 7, our goal was to obtain an expres-
sion for the calibrated flux of M57 in the form of:
  3.2.5 Surface brightness profile
Fcalib
mcalib = −2.5log (9)
texp Another useful tool used is the analysis of the surface
This means that we want to calculate the term that, applied brightness profile. To create this profile for a nebula, its shape
to the flux Finst (5) would give us Fcalib , the calibrated flux. is considered ellipsoidal, which is quite a good approximation.
We would then apply that calculated factor to our final image, The first step is to define various annuli around the centre
for each one of the filters. of the ellipse. The goal is to break down the ellipse into thin
Decomposing equation 6 and simplifying we get: ring like structures, like it is demonstrated in figure 15. The
ellipses are drawn from the centre, with increasing semi-major
k sec z
Finst,0 = 10 2.5 Finst (10) axis. A ring is formed by the thinner pair of ellipses.

8
 light is noticeable.

As mentioned in section 1.1, each color is the footprint of an

element. Oxygen, the most abundant element is present on
the outer ring, followed by nitrogen in slightly lower percent-
age on the outmost ring. The inner "cavity" of the nebula
should, aside from the white dwarf, emit mostly blue from
Helium. Since our acquisitions do not capture this light, we
could not conclude the presence of helium.
Figure 15: Zoomed in image of the nebula M57, using the blue filter,
with elliptical rings.

In each of these annuli, the mean intensity is calculated, as

well as its correspondent error, and in figure 16 we can see the
plot of these mean intensities as a function of the distance to
the centre of the nebula, where each point has associated its
own error bar. It’s possible to see that the intensity is higher
in the centre of the nebula, then it decreases, and it increases
again, around the 70 pixel mark. The intensity in the centre of
the nebula can be explained by the white dwarf there present.
This white dwarf also explains the existence of the nebula
itself. When a star reaches the last stages of its evolution, and
turns into a white dwarf, it sheds its outer layers of ionized
gas out into space, forming the nebula seen [14].
Because M57 has a ring like shape, its contents are con-
Figure 17: Surface brightness profile
centrated around an ellipse with 70 pixels of semi-major axis,
so it is expected that the intensity has a peak around that
mark. On the interior very little light would be emitted if not
for the white dwarf, so the intensity drops, unless on the very
centre where the star is. This describes the curve seen in 16.

3.2.6 Astrometric Calibration (WCS)

To determine accurate celestial coordinates for the images,

astrometric calibration was performed. This process assigns
World Coordinate System (WCS) information to each FITS
image, allowing pixel positions to be translated into right
ascension and declination on the sky. Since the raw tele-
scope images did not include RA and DEC metadata, the
Figure 16: Surface brightness profile for the nebula M57, red filter WCS solution was obtained using the [Link] ser-
vice [15]. The platform automatically detects star patterns
In figure 17 the surface brightness profile is shown for the in the image, matches them with reference catalogues, and
different filters. The intensity of each color has been adjusted returns an updated FITS header containing the astromet-
and calibrated according to equation 12, so that the fluxes ric solution. The solved file can then be opened with the
plotted are all in the same units, and thus can be compared. DS9 program, or with Astropy WCS module, to overlay ce-
The most intense flux is from the green filter, followed by the lestial grids or measure precise coordinates, such as the po-
red, and lastly the blue. The blue filter shows a much lower sition of the central star in M57. In the final RGB photo
flux than the two other filters, and in figure 14 this can be the coordinates obtained for the centre of the image were
seen- the image has very little blue in it. Compared with RA : 18h 53m 32.258s Dec : +33◦ 02′ 09.537′′ . The image
figure 3, an image of the same nebula, the absence of blue scale is: 0.481 arcsec/pixel.

9
 contribution from the surrounding nebular light and to de-
termine a robust estimate of its flux and centroid. Then for
each filter we can correct this flux as we did before to obtain
the calibrated magnitude. The obtained values of magni-
tude for each filter BVR were: 15.405, 15.614, 16.357 cor-
respondingly. Comparing with the tabulated values (table
7) for the magnitudes in each filter obtained in the database
SIMBAD [10] this corresponds to errors of ≈ 0%, 1% and
3%, correspondingly.

Filter SIMBAD Magnitude Data Magnitude ∆m

Figure 18: Main objects in frame identified by [Link]. B 15.405 ± 0.016 15.405 0.000
V 15.769 ± 0.023 15.614 -0.155
R 15.901 ± 0.003 16.357 +0.456

Table 7: Comparison of the central star magnitudes of M57 (B, V, R)

from SIMBAD and our data.

Figure 19: M57 visualized in ds9 program, with WCS coordinate grid

3.2.7 Detection of the Central Star in M57

After the background subtraction and image calibration,

the central star of the Ring Nebula (M57) was identified us- Figure 20: Example of the stars found by DAOStarFinder algorithm in
ing an iterative Point Spread Function (PSF) photometry the V filter.
method.
In this context, the Point Spread Function (PSF) describes
4 Conclusions
how the light from a point source, such as a star, is distributed
on the detector due to the combined effects of the telescope Throughout this work we performed a complete photomet-
optics, the atmosphere, and the CCD response. Instead of ric and astrometric analysis of the planetary nebula M57,
appearing as a perfect point, the star’s light is spread into a from raw CCD data to scientifically meaningful results. The
characteristic profile, typically modeled as a Gaussian func- project integrated multiple stages of astronomical data pro-
tion. This technique allows for precise measurement of stel- cessing, including bias and flat-field correction, alignment,
lar flux and position, even in regions dominated by extended stacking, background estimation, and photometric calibration
emission such as the nebular core. using the nearby standard star Sheliak.
The process begins with an initial source detection per- The data reduction pipeline developed in Python proved
formed by the DAOStarFinder algorithm from the photutils to be efficient in removing instrumental artifacts and back-
package, which locates point sources exceeding a specified ground noise, producing clean and uniform science frames
threshold above the background noise. Among the detected suitable for quantitative analysis. The stacking of multiple
sources, the one closest to the center of the nebula was exposures allowed for an increase in the signal-to-noise ra-
assumed to correspond to the central star. To refine its tio, essential to reveal the structure of the nebula despite the
position and brightness, the detection was passed to the challenging observing conditions of the Lisbon metropolitan
IterativePSFPhotometry routine, which fits an integrated area.
Gaussian PRF (Point Response Function) model iteratively Aperture photometry was employed to measure the flux
to the data. This allows the algorithm to separate the star’s of both M57 and Sheliak in the three BVR filters. Using

10
Sheliak’s catalogued magnitudes, we computed zero-points support and the hospitality of the CIGeoE during the field
that enabled the calibration of the nebular magnitudes, from work at the Observatory. We would like to express our grati-
which a physically accurate RGB composite image was con- tude to all the AL staff for their continuous support through-
structed. The color image shows the characteristic morphol- out this work. A special thanks goes to Professor João Duarte
ogy of the Ring Nebula: an outer red ring dominated by for the guidance and patience that greatly contributed to the
ionized nitrogen and hydrogen, a central greenish region of success of this project. We are also deeply grateful to our
doubly ionized oxygen. However, if fails to properly repre- colleagues, who were always willing to help in the occasional
sent the inner blue region, most likely due to a very low blue moments of despair and made this journey much more enjoy-
count number. The helium that emits this blue color was, able.
therefore, not correctly detected.
The surface brightness profiles revealed the variation of flux
across the nebula, displaying an intensity minimum in the in-
ner cavity followed by a maximum at the main ring radius.
This pattern is consistent with the nebula’s hollow ellipsoidal
structure, where the densest ionized gas is concentrated along
the ring. The peak in intensity around 70 pixels from the cen-
tre corresponds to the nebular shell, while the central emis-
sion can be attributed to the white dwarf remnant.
The astrometric calibration performed with the
[Link] service provided accurate World Coor-
dinate System (WCS) solutions, enabling the translation
between pixel and celestial coordinates. The final image
scale of 0.481 arcsec/pixel was determined, and the nebula’s
central coordinates were measured as References
RA = 18h 53m 32.258s , Dec = +33◦ 02′ 09.537′′ . [1] Center for astrophysics - harvard smithsonian - planetary
nebulas. [Link]
Finally, the central star of M57 was successfully detected
topic/planetary-nebulas. Accessed 10-2025.
and measured using iterative PSF photometry. The method
precisely separated the stellar signal from the surrounding [2] Esa/hubble - planetary nebula. [Link]
nebular emission, yielding calibrated magnitudes in close org/wordbank/planetary-nebula/. Accessed 10-2025.
agreement with the catalogued values from the SIMBAD
[3] Nasa - messier 57. [Link]
database. This confirmed both the accuracy of our reduction
mission/hubble/science/explore-the-night-sky/
pipeline and the validity of our photometric calibration.
hubble-messier-catalog/messier-57/. Accessed
In summary, this work demonstrated the complete obser-
10-2025.
vational and analytical chain required to study extended as-
tronomical objects through CCD imaging. Despite the lim- [4] Alicia Cermak. Decoding nebulae. https:
itations imposed by light pollution and instrumental imper- //[Link]/universe/decoding-nebulae/,
fections, the obtained results are scientifically consistent and August 2021. Accessed: 2025-10-5.
show that small observatories can yield meaningful astrophys-
[5] Planetary Nebulae — [Link]. https:
ical data when supported by proper calibration and compu-
//[Link]/Astronomy/research/
tational analysis. A possible extension of this work would
PN/nebulae/[Link]#:~:text=
be to include infrared observations, enabling the detection of
Astronomers%20are%20drawn%20to%20study,which%
cooler dust and gas that complement the optical emission and
20the%20star%20originally%20formed. [Accessed
offer a deeper understanding of the nebula’s composition and
05-10-2025].
evolution.
[6] Equatorial coordinate system. [Link]
[Link]/cosmos/*/Equatorial+Coordinate+
5 Acknowledgments System. Accessed 10-2025.

This work has been possible thanks to the agreement be- [7] Bradley W. Carroll and Dale A. Ostlie. An Introduction
tween the IST - Instituto Superior Técnico and the CIGeoE- to Modern Astrophysics. Cambridge University Press,
Centro de Informação Geoespacial do Exército. We thank the Cambridge, UK, 2nd edition, 2017 - Chapter 3.

11
 [8] Moravian instruments - ccd cameras. [Link]
[Link]/art?id=321. Accessed 10-2025.

[9] [Link]. [Link] – altitude and airmass calcu-

lator. [Link] Accessed: 2025-10.

[10] Simbad. [Link]

sim-fbasic. [Accessed 10-2025].

[11] Astronomical snr calculator. [Link]

[Link]/sn_calculator_mvn.php. [Ac-
cessed 10-2025].

[12] Ana Mourão. Data reduction. Observational Astro-

physics Course’s Lecture Notes, 2025.

[13] Astropy - creating color rgb images. [Link]

[Link]/en/latest/visualization/[Link].
Accessed 10-2025.

[14] Director’s Office. Planetary nebulas. [Link]

[Link]/research/topic/planetary-nebulas.
Accessed: 2025-10-10.

[15] [Link]. [Link] An online

astrometric calibration service [Accessed 10-2025].

[16] Data analysis code repository. [Link]

Random0utlander/ASTRO. Created 10-2025.

12