Skip to content

Commit 3ed2fc3

Browse files
authored
DOC: EHT Case Study Suggestions (#196)
* DOC: Re-work EHT intro paragraph * Removed sentence about EHT sensing wavelength. * Updated wording surrounding VLBI * Switch to md reflink style for source readability * Updates to Key Imaging Objectives * Renamed the section to 'Key Goals and Results' to better fit the content. * Updated first bullet point with context and accurate date. * Minor re-wording of remaining bullet points * Modified the wording of a couple bullet points. * Made the reference to M87's black hole more explicit * Updated last bullet point of Challenges to give a higher-level overview of the imaging challeng. * Re-wrote first paragraph of NumPy's role section * Attempting to focus specifically on image reconstruction, particularly eht-imaging, where the dependence on the scientific Python ecosystem is most relevant. * Trying to provide a little background on overall reconstruction task, and tie it into why NumPy/SciPy are so important. * More reorganization of the NumPy Role section. * Moved eht-imaging up as it ties in with the theme and is the most prominent example of NumPy used in the eht analysis * Removed paragraph on CHIRP: - Trying to reduced redundancy given previous discussion on data processing - CHIRP is not available on GH - not sure to what extent NumPy is used for that particular pkg * Removed additional paragraph on regularized MLEM - want to avoid too much detail and focusing on algorithms * Re-wording of the summary. * replaced vectors/matrices with n-dimensional array * Emphasize importance of discovery and role of collaboration. * Final pass through for wording/grammatical edits. * Added bolded titles to bulleted list in eht casestudy. Addresses comments from review.
1 parent ad8334d commit 3ed2fc3

File tree

1 file changed

+77
-74
lines changed

1 file changed

+77
-74
lines changed

content/en/case-studies/blackhole-image.md

+77-74
Original file line numberDiff line numberDiff line change
@@ -10,39 +10,46 @@ sidebar: false
1010
<footer align="right">—Katie Bouman, <cite>Assistant Professor, Computing & Mathematical Sciences, Caltech</cite></footer>
1111
</blockquote>
1212

13-
## About Event Horizon Telescope
13+
## About The Event Horizon Telescope
1414

1515
The [Event Horizon telescope (EHT)](https://eventhorizontelescope.org), is an
1616
array of eight ground-based radio telescopes forming a computational telescope
17-
the size of the earth, that are designed to study extreme objects in the
18-
universe, with unprecedented sensitivity and resolution. It is a worldwide
19-
network of eight pre-existing telescopes based on a technique called
20-
very-long-baseline interferometry (VLBI). This technique is used to
21-
synchronize these telescopes deployed across the globe to form one huge,
22-
Earth-size telescope capable of observing at a wavelength of 1.3 mm. This
23-
means EHT can achieve an angular resolution of
24-
[20 micro-arcseconds](https://eventhorizontelescope.org/press-release-april-10-2019-astronomers-capture-first-image-black-hole)
25-
enough to read a newspaper in New York from a sidewalk café in Paris!
26-
27-
### Key Imaging Objectives
28-
29-
* To study the most extreme objects in the Universe predicted by Einstein’s
30-
theory of general relativity, during the centennial year of the historic
31-
experiment that first confirmed the theory (2017).
32-
33-
* Focus on the [black hole](https://solarsystem.nasa.gov/resources/2319/first-image-of-a-black-hole/)
34-
at the center of Messier 87 galaxy, located in the Virgo galaxy cluster.
17+
the size of the earth, designed to study extreme objects in the
18+
universe with unprecedented sensitivity and resolution. The worldwide
19+
network of radio telescopes comprises a virtual telescope based on a technique
20+
called very-long-baseline interferometry (VLBI).
21+
Using this technique, the EHT is able to achieve an angular resolution of
22+
[20 micro-arcseconds][resolution] — enough to read a newspaper in New York
23+
from a sidewalk café in Paris!
24+
25+
[resolution]: https://eventhorizontelescope.org/press-release-april-10-2019-astronomers-capture-first-image-black-hole
26+
27+
### Key Goals and Results
28+
29+
* **A New View of the Universe:**
30+
The EHT is an exciting new tool for studying the most extreme objects in the
31+
universe. The EHT's groundbreaking image was published 100 years
32+
after [Sir Arthur Eddington's expidition][eddington] yielded the first
33+
observational evidence in support of Einstein's theory of general relativity.
34+
35+
* **Investigating Black Holes:**
36+
The EHT's first image focuses on the supermassive black hole at the center
37+
of the galaxy Messier 87 (M87), located in the Virgo galaxy cluster.
3538
This black hole resides approximately 55 million light-years from Earth and
3639
has a mass equal to 6.5 billion times that of the Sun. It has been a
3740
subject of astronomical study for
3841
[over a 100 years](https://www.jpl.nasa.gov/news/news.php?feature=7385).
39-
Black holes have been theoretically predicted and observed but a real image
40-
was never created until now.
42+
Black holes have long been the object of intense study but the EHT provides
43+
the first direct visual evidence of these extreme phenomena.
4144

42-
* Based on Einstein’s general theory of relativity, the scientists expected to
45+
* **Comparing Observations to Theory:**
46+
Based on Einstein’s general theory of relativity, scientists expected
4347
see a dark region similar to a shadow, caused by the gravitational bending
4448
and capture of light by the event horizon. By studying this shadow
45-
scientists could measure the enormous mass of M87’s black hole.
49+
scientists could measure the enormous mass of M87’s central supermassive
50+
black hole.
51+
52+
[eddington]: https://en.wikipedia.org/wiki/Eddington_experiment
4653

4754
### The Challenges
4855

@@ -60,75 +67,71 @@ enough to read a newspaper in New York from a sidewalk café in Paris!
6067
Terabytes worth of observed data per day, stored on high-performance
6168
helium filled hard drives.
6269

63-
* **Source Imaging and Model Fitting**
70+
* **Image Reconstruction**
6471

65-
The sequence of correlation and engineering releases represents a
66-
year-long effort of identifying and mitigating data issues, and developing
67-
new software and procedures that could reliably choose the most likely
68-
image based on actual measurements.
72+
How are the calibrated data processed to produce an image of something that
73+
has never before been directly imaged? How can scientists be confident
74+
that the image is correct? These are some of the challenges overcome in
75+
the analysis to produce the image.
6976

7077
{{< figure src="/images/content_images/cs/dataprocessbh.png" class="csfigcaption" caption="**EHT Data Processing Pipeline**" alt="data pipeline" align="middle" attr="(Diagram Credits: The Astrophysical Journal, Event Horizon Telescope Collaboration)" attrlink="https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57" >}}
7178

7279
## NumPy’s Role in Black Hole Imaging
7380

74-
There are several aspects to black hole imaging besides data collection, noise
75-
elimination, data cleanup, reduction and correlation. Imaging is crucial as it
76-
can help to predict not only the black hole mass but also rule out whether a
77-
black hole could be a wormhole, a theoretical bridge between distant points
78-
in spacetime. But it is also incredibly hard to measure given the astronomical
79-
distances involved. As Katie Bouman mentions in her
80-
[TED talk](https://www.youtube.com/watch?v=BIvezCVcsYs),
81-
‘It is like taking a picture of an orange on the surface of the moon.’
81+
While collecting, curating, and processing the data from the EHT facilities
82+
represents a monumental challenge, it is only the first step in generating
83+
an image from the data.
84+
There are many approaches to image reconstruction, each incorporating unique
85+
assumptions and constraints in order to solve the ill-posed problem of
86+
recovering an image of the black hole from the collected data.
87+
But how can anyone be confident that the image that's produced is correct?
88+
What if there's a problem with the data? Or perhaps an algorithm relies too
89+
heavily on a particular assumption? Will the image change drastically if a
90+
single parameter is changed?
91+
The EHT collaboration met these challenges by having independent teams
92+
evaluate the data using both established and cutting-edge image reconstruction
93+
techniques to verify that the resulting images were consistent.
94+
Results from these independent teams of researchers were combined to yield the
95+
first-of-a-kind image of the black hole.
96+
This approach is a powerful example of the importance of reproducibility and
97+
collaboration to modern scientific discovery, and illustrates the role that
98+
the scientific Python ecosystem plays in supporting scientific advancement
99+
through collaborative data analysis.
82100

83101
{{< figure src="/images/content_images/cs/bh_numpy_role.png" class="fig-center" alt="role of numpy" caption="**The role of NumPy in Black Hole imaging**" >}}
84102

85-
Once the key challenges posed by EHT, data collection and reduction are taken
86-
care of, the next big challenge in data processing is related to imaging. The
87-
imaging algorithms form the core of this task as through imaging, scientists
88-
could calculate the shadow of the black hole which forms the crux of several
89-
other calculations related to event horizon and nearby objects. One of the key
90-
algorithms used in imaging was developed by Katie Bouman – Continuous
91-
High-resolution Image Reconstruction using Patch priors, or ‘CHIRP’. It can
92-
parse the cumulative telescope data gathered by the Event Horizon Telescope
93-
project.For imaging tasks, researchers banked on Python to run the datasets on
94-
these algorithms, arraying and plotting data for meaningful insights.
95-
96-
Besides NumPy, there were other packages such as
97-
[SciPy](https://www.scipy.org), [Pandas](https://pandas.io) and
98-
[Matplotlib](https://matplotlib.org) that helped in imaging and
99-
[data processing for imaging the black hole](https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57).
100-
The standard astronomical file formats and time/coordinate transformations
101-
were handled by [Astropy](https://www.astropy.org) while Matplotlib was used
102-
in visualizing data throughout the analysis pipeline, including the generation
103-
of the final image of the black hole.
104-
105-
Andrew Chael and the ehtim project team came up with
106-
[eht-imaging](https://github.com/achael/eht-imaging) Python modules for
107-
simulating and manipulating VLBI data and producing images with regularized
108-
maximum likelihood methods. NumPy is at the core of array data processing used
109-
in this software package named ehtim as indicated by the partial software
103+
For example, the [`eht-imaging`][ehtim] Python package provides tools for
104+
simulating and performing image reconstruction on VLBI data.
105+
NumPy is at the core of array data processing used
106+
in this package as illustrated by the partial software
110107
dependency chart below.
111108

112109
{{< figure src="/images/content_images/cs/ehtim_numpy.png" class="fig-center" alt="ehtim dependency map highlighting numpy" caption="**Software dependency chart of ehtim package highlighting NumPy**" >}}
113110

114-
The challenge posed during reconstruction of an image using VLBI measurements
115-
is that there can be an infinite number of possible images that explain the
116-
data. The ehtim software addresses this challenge by implementing algorithms
117-
that help find a set of most likely reasonable images that respects prior
118-
scientific assumptions while still satisfying the observed data.
111+
[ehtim]: https://github.com/achael/eht-imaging
112+
113+
Besides NumPy, many other packages such as
114+
[SciPy](https://www.scipy.org) and [Pandas](https://pandas.io) were used in the
115+
data processing pipeline for imaging the black hole.
116+
The standard astronomical file formats and time/coordinate transformations
117+
were handled by [Astropy][astropy] while [Matplotlib][mpl] was used
118+
in visualizing data throughout the analysis pipeline, including the generation
119+
of the final image of the black hole.
119120

121+
[astropy]: https://www.astropy.org/
122+
[mpl]: https://matplotlib.org/
120123

121124
## Summary
122125

123126
NumPy enabled researchers to manipulate large numerical datasets through its
124-
efficient data structures of vectors and matrices leading to imaging and
125-
plotting of the first ever image of a black hole. Imaging of M87 black hole is
126-
a major scientific feat that was almost presumed impossible a century ago. In
127-
a way, black hole image has helped in a stunning confirmation of Einstein’s
128-
general theory of relativity. This is not only a breakthrough in technology,
129-
but an example of international scale collaboration that uses connections
130-
between the world's best radio observatories, over 200 scientists worked with
131-
observations collected over 10 days and analyzed for over a year. They used
127+
efficient and generic n-dimensional array, providing a foundation for the
128+
software used to generated the first ever image of
129+
a black hole. The direct imaging of a black hole is
130+
a major scientific accomplishment providing stunning, visual evidence of Einstein’s
131+
general theory of relativity. This achievement encompasses not only
132+
technological breakthroughs, but international-scale scientific collaboration
133+
between over 200 scientists and some of the world's best radio observatories.
134+
They used
132135
innovative algorithms and data processing techniques improving upon existing
133136
astronomical models to help unfold some of the mysteries of the universe.
134137

0 commit comments

Comments
 (0)