No Halleymites in the Main Belt!

Today we have a guest post by Comet Hunters volunteer Peter Jalowiczor.

Peter Jalowiczor (left) with MSAS Chairperson Les Marsden (right)

I enjoy trawling through quality old vinyl to buy, and sometimes real gems can be found such the one I came across recently to my amazement (and some amusement): a 7“ vinyl single commemorating the return of Halley’s Comet in 1986 with a children’s book feel to the cover…

Yes, we’re real Halleymites from the Comet Halley and Yes, You’ve never heard of us before but we’ve been around the Galaxy for thousands of years…

How prophetic. I certainly have never heard of them. At the time Comet Halley came around in 1986, I was a first year Physics Student at the University of Sheffield. This astronomical event was big news. A few years later, I would engage in a slice of Postgraduate research for P/ Halley  under one of the academic staff who was a World expert in his field at the time.

But the Halleymites somehow kept quiet…

The Talk

Over thirty years later and we have Comet Hunters; it is a project I have contributed towards alongside other Zooniverse projects (mainly P4). By classifying images whenever there were a few minutes. Recently, I had the opportunity of collating materials from Comet Hunters for a talk to be given at my local Astronomical Society: the MSAS (the Mexborough and Swinton Astronomical Society). The society is situated ~20km from Sheffield (pop, 570,000), England and was founded in 1978.  Every Thursday evening is a social occasion centred on a lecture.  Members, such as me are encouraged to give talks. Usually about once a month, an academic visits the society to present on an aspect of Astronomy.

The evening was divided-up into four parts:

  1. Q & A from the Planet Four talks at MSAS
  2. Presentation – based largely on the blogs Peter Jalowiczor
  3. Presentation – material kindly provided by Meg Schwamb
  4. Q & A on the Comet Hunters’ Presentations

Q & A from the Planet Four talks at MSAS

This was a chance to review the Q & A to make sure anyone, who had missed these questions posted on the P4 blog

II.& III.    The Presentations

Started with an introduction to Comet Hunters as a citizen science project. The project was initiated to greatly increase the discovery rate of these objects. The first such object was found in 2006 and there are only 12 known at present.

This followed with a discussion about Asteroids and Comets with the distinct examples of Gaspra and Halley being given in these categories. However, where do we draw the distinction between these objects? Comets were always assumed to have come from the outer Solar System beyond the orbit of Neptune. Asteroids have generally thought to be from the Main Belt.

Next came a discussion of the two different searches: the HSC Survey or the Archival Data Search.  A discussion of the instruments: Suprime Cam and the Hyper Suprime-Cam used in the search for these objects.

The Subaru telescope – this also included Meg’s detailed description about life for Astronomers at this wonderful observatory; one of the best observing sites in the World. The difficulties experienced by Astronomers at such high altitudes over 4,600m asl. Anyone working at the summit of Mauna Kea cannot stay up longer there longer than 14 hours needing to descend to mid-altitude for safety and health reasons. Sandwiches left for more than 20 minutes become stale giving all the water to the air (as the one left by Meg). This prompted one of the members to jokingly suggest Rhyita might be a better option next time.  Another of the society members (not present) has travelled the world specifically visiting observatories and had the privilege of visiting the observatory in 2012 at the time of the transit of Venus.

Conclusions

  • Hyper Suprime-Cam search launched in June 2016.
  • 0ver 8,000 volunteers have contributed to the project to varying degrees, inspecting HSC for main belt comets.
  • 4,877 main belt asteroids located in the HSC wide survey fields have been searched for cometary activity.
  • Assuming a detection efficiency of 20% an upper limit of 1% is placed for the occurrence rate of MBCs in each of the sample of asteroids.
  • Simulated MBC observations are planned to measure the detection efficiency.

Q & A

A variety of questions were asked following the two presentations. The Comet Hunters science team will try and answer these questions in a following blog post.

  • Snow Line – how exactly is the boundary of the snow-line worked out? What is its relevance? It seems that there is water (albeit in different forms) such as lots here on the Earth both inside and outside the snowline?

 

  • Regarding the MBCs isn’t there a possibility that the surface ices over geological time would have sublimated from the outside (of the object)? We would then be left with ice at the centre perhaps covered with rock/dust?

 

  • Classification: What is the point of the reference images: the ones to the right of the main image when we are really looking at the central object?

 

  • What is the file size from the updated camera (i.e. Hyper Suprime-Cam)?

 

  • In the charts showing the ‘Tail Visible in One of the Images’ and ‘Tail Visible in Two of the Images’ what exactly does this data represent?

The presentations certainly opened up some debate during and after the talks. A number of members noted that if we look at the current definition of Comets and Asteroids, there does not seem to be a clear boundary.  How would Enceladus fit in this category? After all, it is ejecting water.

On an ending note, I also think that it is safe to say that there are no Halleymites in the Main Belt…

Audience at the MSAS meeting

 

New Archival Search Images

A quick note to announce new archival search images on the site. We’re gotten through about two thirds of the processed Suprime-Cam images from  9,850 exposures (or 9,850 exposures  x 10 chips on Suprime-cam = 98,500 FITs files!) from the date between 2007 Jan to 2013 Feb.  These are observations where the exposure time is less than 60 seconds. With more data, there’s new asteroids to search, giving new opportunities to find the elusive main belt comet. If you don’t find a main belt comet, your classifications will help us put limits on the frequency of these types of outbursts in the asteroid’s main belt, which is still important. Dive in today at http://www.comethunters.org

 

New HSC Images On Comet Hunters

We have new HSC images available on Comet Hunters from last August now available now on the site. With more asteroids, there are more chances to identify cometary activity. If you don’t spot a tail, that’s okay too. You’re helping us figure out how frequent these cometary outbursts are.

Also today, the Hyper Suprime-Cam Survey, which we get our asteroid images from, just had their first public data release. You can read more about it, and see some stunning images from the HSC camera here.

Happy Comet Hunting!

Making a Push on the Suprime-Cam Search and Waiting on New HSC Images

I wanted to give an update on both Comet Hunters Searches

HSC Search: Thanks to your help, we’ve completed all the live HSC images. We’re currently working on processing more images. We had some data processing challenges that are not solved. We hope to get new images on site by the end of February. Stay tuned for to this space for more updates.

Archival Search: We’re working towards the first paper, that will focus on the Suprime-Cam Archival Search. We’ve started to work on some of the paper text and analysis. One of the next steps is to compare to what automated analysis suggests might have a point-spread function.  We think this would be an interesting comparison. We’d like to include as much completed Suprime-Cam observations in our analysis as possible. If  you can spare some time, please classify an image or two on the Archival Search today at http://www.comethunters.org

Previous Trailed MBC Discoveries

Because of the longer exposure times used by the Hyper Suprime-Cam survey, the asteroids look bean-shaped or elongated in the HSC search. This trailing is due to the asteroid’s relative motion to the Earth. We’re seeing it move and therefore its light is deposited in slightly different positions along the camera’s field-of-view during the exposure, creating a streak or trail. As far as we can find in the scientific literature, nearly all main-belt comets were detected in images will little trailing. We found one example, that we wanted to share with you. Below  is a well-known image of  Main Belt Comet 107P/Wilson-Harrington. This observation was actually taken back in 1949 (on glass plates!) by the 48-inch Oschin Schmidt telescope on Mt. Palomar for use in the Palomar Observatory Sky Survey (Fun Fact – I used the same telescope for my thesis survey for distant Kuiper belt objects nearly 60 years later).

Yanga  Fernandez, Lucy McFadden, Carey Lisse, Eleanor Helin, and Alan  Chamberlin went back to these observations in 1996/1997 and found that this asteroid was active as an MBC with a visible tail! You can find their paper here (unfortunately it’s behind a paywall, but you can read the paper abstract for free). The  asteroid is very streaked, a bit more than what you typically see on the Comet Hunters HSC search, but it gives you an idea. The tail is faint and diffuse, but visible off to the left of the streak.

In case you need help spotting the tail, I’ve annotated this version below with magenta arrows pointing to the tail.

f02513dc-bf36-49ec-9fbe-cba7f417fa3a

Adapted from DSS/ Fernandez et al., 1997, Icarus, Vol. 128, pg. 114-126. Magenta arrows indicate the faint tail.

 

The Shape of an Asteroid

I wrote this sitting on a plane on the way to Grapevine, Texas. I’m on my way to the American Astronomical Society winter meeting. It is the largest meeting of astronomers in the United States.  I’ll be presenting a talk on Comet Hunters about the Hyper Suprime-Cam search.

A question that pops up from time to time on Comet Hunters Talk is whether or not we can see the shape of an asteroid in the HSC and Suprime-Cam observations. The answer to that is no. The Subaru Telescope does not have the resolution to resolve the shape of the asteroid. The asteroid is viewed by the telescopes and cameras as  point-like and smeared out to the turbulence in the atmosphere and the optics of the cameras just like the background stars, so we can’t infer anything about the shape of the asteroid in most case just from what we see in the Comet Hunters images displayed on the site. In the HSC workflow, many of the asteroids look streaked or ‘bean shaped’ compared to the stationary background stars. This is due to the asteroid’s motion. The HSC observations are close to 150 seconds, and in that amount of time some asteroid orbits have on-sky velocities that move a noticeable number of pixels elongating the asteroid’s appearance in the image.

bb309373-b0ba-49bd-8b7e-78dd9b99d6f5

Very fast moving asteroid in two HSC observations

But there are other ways to indirectly probe the shape of the asteroid. You can use the varying amount of light reflected by the asteroid over time to estimate  the shape of the asteroid and how it rotates. Asteroids don’t produce their own light source. In the optical wavelengths (what our eyes can see), asteroids are reflecting a portion of the Sun’s light. How much surface area and the type of surface changes the amount of sunlight reflected back to the Earth. If the object is very round, you’ll see a nearly uniform amount of light from the object. If the asteroid is oblong, you’ll see the object brighter when the longer axis is facing Earth and a see it is fainter when the smaller axis as the body rotates. This picture can be complicated if there is compositional differences on the asteroid’s surface and how much light those different surface types absorb and reflect sunlight. I high recommend checking out Pedro Lacerda’s  light curves of small solar system bodies website to see simulated light curves (brightness measurements over time) of small solar system bodies.

Happy 1st Birthday Comet Hunters!

Image credit Will Clayton – flikr

Today marks one full year of Comet Hunters. It’s amazing how time flies while you’re having fun and exploring the Solar System! Thank you for all of your time and  contributions. We really appreciate it. We couldn’t do this without the Comet Hunters volunteer community.

Over this year, the science team has been developing the analysis pipeline to combine the multiple classifications of each subject together. We’ve also reviewed possible comet candidates and also  launched the Hyper Suprime-Cam search. In Year 2, we’ll work towards  the project’s first science papers. I’ll be marking Comet Hunters’ birthday by sitting down in a coffee shop and writing some text about how the Comet Hunters website works for the start of our first science paper draft. Stay tuned to this space in the future for more paper and progress updates.

Help celebrate Comet Hunters’ first birthday by classifying an image or two at http://www.comethunters.org

Help Build a Better Zooniverse while Searching for Main-Belt Comets

Starting today you might see some new things popping up on Comet Hunters. These new features and surveys have been developed by the Zooniverse that may one day end up on all future new Zooniverse projects. The Comet Hunters  community was selected to help to try these features out. During the next four weeks, selected Comet Hunters volunteers will receive an invitation to participate as you are classifying images in the HSC search workflow.  If you just want to get on with the comet search, not to worry. You can opt out of participating or check out the Archival search which is exactly the same as before.

Thanks in advance for your help. Your feedback and interactions will help the Zooniverse to improve the current set of web tools and  build better future Zooniverse projects.

New Asteroids to Review on Comet Hunters

We’ve recently  uploaded brand new images on both the HSC workflow and the Archival Workflow.

The Archival images are still from Suprime-Cam and  continuing to move backward in time. These observations will help us understand the true frequency of main-belt comets. Any good candidates we’ll follow-up when the asteroids return to the same spot in their orbit as when the observation was taken to see if the activity repeats.

We’ve now uploaded June HSC observations on to the HSC workflow. These images are  hot off the telescope, giving us a chance to follow-up these asteroids now if they are active.

Both sets of  images are of brand new asteroids never viewed before on the site with  new chances to discover a  main-belt comet. You might be the first person to know a new comet is lurking in our Solar System’s asteroid belt. Search for main-belt comets today at http://www.comethunters.org!

A Comet Hunters Summer

Ishan has spent the last two months as an ASIAA summer student with Comet Hunters working on making simulated images of main-belt comets. Ishan has chronciled his progress here and here on the blog. The goal is that with this simulated main-belt comets we can learn about how faint of cometary activity can the project see. This will let us probe the frequencies of main-belt comets in ways we couldn’t do with just the HSC data alone. Thanks Ishan for all your hard work and effort this summer.

Ishan gave his final presentation at the end of August which he is kindly sharing here:

grouppresentation_mod-001Hello I am Ishan Mishra and my project is ‘Studying Main Belt Comets with Comet Hunters’

grouppresentation_mod-002Let’s quickly review what are the major differences b/w comets and asteroids:

  1. As you can see in this graphic, most of the comets reside in the Kuiper belt beyond the orbit of Neptune and in the Oort Cloud in the outer solar system.
  2. The also differ in size and composition. While comets typically range from 6 – 25 miles in diameter, asteroids are much larger with their diameters ranging from the size of small rocks to about 600 miles.
  3. Comets contain a lot of ice along with rock and hydrocarbons while asteroids are composed of rocks and metals

grouppresentation_mod-003So what are Main-Belt Comets? These are a recently discovered class of main belt objects which show cometary activity. Here’s an image of a main belt comet taken by the Subaru Telescope.

grouppresentation_mod-004Here we see images of the first six MBCs discovered. The plot on the right clearly shows that these objects are distinct from comets while being indistinguishable from asteroids.

grouppresentation_mod-005

So why are we interested in MBCs?

  1. Firstly, our classic difference between asteroids and comets based on their composition is called into question. Asteroids were never known to have volatiles present in them.
  1. Present-day sublimation in the main belt is surprising, as surface ice is unsustainable over the age of the Solar System so close to the Sun. MBCs provide new constraints to the location of the so-called snow-line (the distance from the Sun beyond which primordial temperatures in the protoplanetary disk were low enough for water to condense as ice and thus be swept up into forming planetesimals)
  1. The astrobiological implication: Origin of water on Earth. This interests me the most!

The Earth is believed to have formed dry owing to its location inside the snow line, and therefore the water we see today must have been delivered from outside the snow line, perhaps by impacting asteroids and comets. There are models which suggest that water delivery from asteroid belt was more likely. MBCS provide us a way to test this.

grouppresentation_mod-006However, so far very few MBCs, about 12, have been discovered. We need to find more in order to:

  1. Determine their extent and abundance
  2. Having a large dataset will help us understand them better

The search efforts so far have been mainly of two kinds: Using computer based pattern recognition algorithms and using human pattern recognition abilities.

  1. Computer algorithms currently are not faring well because of the very low SNR of the coma/tail region. They are also unable to accurately account for the whole range of tail morphologies. Hence, there are a lot of false positives.
  1. Telescopic surveys provide the large datasets needed to increase chances of finding objects like MBCs. PAN-STARRS is one such survey which uses a 2 m telescope in Hawaii. Here is an image of the first MBC discovered using PAN-STARS dataset. There are a couple of issues with this method though:
  1. The image quality is not good as its a small telescope. In fast, most surveys use a small telescope.
  2. Since the search is conducted by eye and the team consists of just a few scientists, the process is very slow and laborious.

Here’s where Comet Hunters comes in as a possible savior.

grouppresentation_mod-007So, Comet Hunters is a Citizen Science project under the citizen science web portal called Zooniverse. The idea behind citizen science is crowdsourced scientific research, ie, a large number of volunteers, their numbers usually in thousands, actively participating to complete research tasks.

Projects at Zooniverse span Astronomy, Climate Science, Biology, Physics and even Humanities. Galaxy Zoo is a famous example of astronomy related. The volunteers are shown images of galaxies and they have to classify them according to their morphologies. Another is Planet Hunters, where you need to spot dips in light curves corresponding to planet transition.

In Comet Hunters, volunteers are shown images of asteroids and they need to check for the presence of a coma or a tail. A candidate is vetted by the research team only when about 20 volunteers spot any activity. The combined assessments of so many people are much more reliable than individual opinions.

  1. So we have a large number of searchers now. What about the data?

grouppresentation_mod-008The data for Comet Hunters comes from Subaru, an 8 m telescope at Mauna Kea, Hawaii.

Two key advantages:

  1. Large telescopes are rarely used for surveys. So we have a large amount of high quality data.
  2. Two catalogs: Suprime Cam Archive and the Hyper Suprime Cam current survey data. So we have 17 years worth of data. Easier to spot recurrent activity.

grouppresentation_mod-009Here’s a just a quick example of an image shown on the Comet Hunters website.

It can be also be seen with inverted colors for more clarity.

The two images are about 20 arcsec in width each and are from the same night.

The volunteers need to answer whether a tail is present in one image, both images or is completely absent.

  1. We have a large number of people now working on high quality data. But,

What is the detection efficiency of Comet Hunters? How good is the project in detecting MBCs? What kind of MBCs are getting detected?

This where my project comes in.

grouppresentation_mod-010My job is to create or simulate asteroid images similar to Subaru’s, with varying properties like the coma or tail strength, tail direction, etc.

The volunteers’ choices for these images will help us characterize the kind of MBCs that are being detected by Comet Hunters. For example, we can find the minimum strength of the coma/tail for which the volunteers are able to spot them.

Here are a two examples of images that I have generated till now. I will explain the process in the subsequent slides.

grouppresentation_mod-011

Before I start talking about my work, I need to give some background information.

  1. I will be using this term called ETA to denote the strength of coma activity. As you can see, a value of 0 means absence of coma.
  1. If you look closely in the Subaru images, you will find that the asteroid image is elliptical not circular. I exaggerated the stretching in the model images I displayed in the previous slide. And you can also see this in on the image on left.

So why does this happen?

Solar system objects move much faster in the plane of sky than the distant stars. And Subaru’s survey data images are usually taken for exposure times of 120-150 seconds, as they are not specifically designed for solar system observation.

So during this interval the asteroids/comets move a couple of arcsec resulting in smearing or stretching of the image in the direction of motion.

grouppresentation_mod-012Here I present a simplified pipeline for the generation of model MBC images. As you can see, there are a lot of steps involved. Let’s go through them one at a time.

grouppresentation_mod-013

  1. The first part is the generation of a simple asteroid image with or without a coma, whose intensity strength we can control.
  2. The second box says high resolution asteroid image. This means that we are starting with images at a much higher resolution than that of images obtained from Subaru. I will come to the reason shortly. (*Planning to talk about it when I explain bin and sum*)

Let’s see how a simple asteroid looks.

grouppresentation_mod-014Here is our asteroid, just a pixel wide, with no coma. So, all the pixels in this image have value zero except the central pixel which has the value 1.0.

  1. This is how the asteroid will look if we see it in space, without any optical or atmospheric effect. The ideal case.
  2. The image is 20 arcsec in width with a resolution of 0.01 arcsec per pixel (compared to Subaru’s resolution which is approx. 0.17 arcsec per pixel).
  3. Let’s zoom in to the central 1 arcsec region to see the asteroid clearly

grouppresentation_mod-015So we see a tiny dot at the center which represents the asteroid.

grouppresentation_mod-016

  1. Now let’s add a coma to the asteroid. If you remember, we defined a parameter ETA which is proportional to the intensity strength of the coma. Here you see an asteroid with a coma of intensity 1.0
  2. The coma is invisible in the default linear scale image. So I have also shown the log scale image to confirm the presence of a coma.

grouppresentation_mod-017

  1. Here are some images of asteroid with coma of different intensities. The ETA increases clockwise. Again, these images are all in log scale for clarity.

So, we have our asteroid image with a coma around it. Let’s move on to the next stage.

grouppresentation_mod-018

  1. We now trail the asteroid. As I had mentioned earlier, trailing is necessary to account for the stretched asteroid image which is due to the significant movement of asteroid in sky during the telescope exposure time.
  1. The information we need to trail the asteroid is its speed and direction of movement as seen in the plane of the sky and the exposure time of the telescope, which will control the length of the trailed image.

Let’s see an example.

grouppresentation_mod-019

  1. Here I trail an asteroid with coma of ETA value 25.0. We trail it with a speed of 50 arcsec/hour, at an angle of 45 degrees for an exposure time of 60 seconds.
  1. I have purposely shown the asteroid image in linear scale so that we can see the difference in the background intensity between the two images. Naturally, the comae from adjacent asteroids add up in the second image to make it look brighter.

Let’s zoom in.

grouppresentation_mod-0201. We zoom into the central 1 arcsec region and now we can clearly see the individual asteroid replicas.

2. The gap between successive asteroids is because we are imaging in an interval of 0.5 seconds for the duration of exposure. Hence, the asteroid is displaced by the distance it moves in this small interval.

grouppresentation_mod-021

Just a quick look at the trailed asteroid images for different coma intensities. All these images are of the central 1 arcsec region and are shown in logscale.

OK. Now that we have the trailed asteroid, we proceed to a very important step: Reducing the resolution or increasing the pixel scale to Subaru’s.

grouppresentation_mod-022A look at the trailed asteroid images for different trailing directions

grouppresentation_mod-023

  1. So in this step we go concert our high resolution images, of pixel scale 0.01 arcsec/pixel, to the lower Subaru resolution of 0.17 arcsec per pixel. But why did we start with a higher resolution?
  1. We are starting with how the image looks from space, before passing through the asteroid and being images by the CCD in the telescope. This is why we started with a single pixel wide asteroid, a point source.
  1. So, we are basically replicating how a CCD works, with all the photons falling in a square CCD pixel account for its value.

grouppresentation_mod-024

  1. So in this step we go concert our high resolution images, of pixel scale 0.01 arcsec/pixel, to the lower Subaru resolution of 0.17 arcsec per pixel. But why did we start with a higher resolution?
  1. We are starting with how the image looks from space, before passing through the asteroid and being images by the CCD in the telescope. This is why we started with a single pixel wide asteroid, a point source.
  1. So, we are basically replicating how a CCD works, with all the photons falling in a square CCD pixel account for its value.

grouppresentation_mod-025Here’s a practical example: the famous Eagle nebula.

Before we proceed, I would like to mention that prior to using this binning and summing technique to decrease resolution, we were trying to use interpolation. It didn’t work out at the end as we discovered a major flaw with the method or maybe the interpolation function we were using wasn’t working as expected. We can talk about this later if anyone’s interested.

OK. So let’s see how the low resolution images of our trailed asteroid look.

grouppresentation_mod-026Here’s a high resolution trailed image that we saw before. The input parameters are also shown. As you can see, this one has ETA = 0 , ie , no coma.grouppresentation_mod-027This is the low resolution image. It is 101 x 101 pixels wide. Please note that I have verified the total flux (in this case the sum of all pixel values) is conserved during the sum and bin process.

grouppresentation_mod-028

Here are some more low resolution image with same parameters as the previous image except ETA.grouppresentation_mod-029OK. Let’s move on now to the final stage: convolving our model asteroid with a background star. But why do we need this?grouppresentation_mod-030To account for two things that affect the ideal image: Atmosphere and Telescope optics. Convolving with a background star incorporates these effects into our ideal image and makes it looks realistic. grouppresentation_mod-031Here’s the background star which we will use for the convolution process. This was chosen by eye from a field image obtained from HSC database.

You will notice that the image on the right here and many subsequent images are shown in zscale. This is because the Subaru images fed into the CometHunters website are in zscale. This scaling provides a convenient perspective to look for any activity in the object.

grouppresentation_mod-032

Let’s see how our convolution process fares for some model images we have seen so far.

Case 1: The most basic one: untrailed asteroid with no coma.

As you see, since the model asteroid in this case is basically a delta function, the background star image is identically reproduced.

grouppresentation_mod-033

Case 2: Horizontally trailed coma asteroid with no coma.

Now here we encounter a major problem. The noisy background from the star’s image is getting trailed in the direction of asteroid’s movement. We have been trying to remedy this by playing with the convolution process but to no avail.

grouppresentation_mod-034

Recently, we came up with a neat trick. Why not generate a trailed sky background image same as in the convolved output but without the central asteroid. Then, ideally, subtracting this trailed background with our convolved output should remove the background from the latter image, leaving just the asteroid part we are interested in.

Let’s see how this works.

grouppresentation_mod-035

Firstly, to generate the sky background we need to convolve our model asteroid with the background image (with star removed). We remove the star from the background image by simply replacing a central square region containing the star (the image on slide 31) with some random background sky region. This process is crude and not the best way to remove the star, but will have to work for now.

You can see that the background sky gets trailed when convolved, as expected.

grouppresentation_mod-036

Now we subtract our convolved output (from slide 33) with our trailed background image from the previous slide and…..Voila! The background gets removed perfectly!

grouppresentation_mod-037

Now the final step is to add noise in this image. We add poisson noise to each pixel because of something called shot noise.

 

Shot noise is caused by the random arrival of photons.  This is a fundamental trait of light.  Since each photon is an independent event, the arrival of any given photon at a pixel cannot be precisely predicted; instead the probability of its arrival in a given time period is governed by a Poisson distribution.

Slide 38:

 

Here you look at our final output image. Looks good, doesn’t it?!

grouppresentation_mod-038grouppresentation_mod-039

Some sample cases

grouppresentation_mod-040

grouppresentation_mod-041grouppresentation_mod-042grouppresentation_mod-043grouppresentation_mod-044grouppresentation_mod-045grouppresentation_mod-046

Conclusion and Future work.
Although the pipeline is not completely ready to generate a the complete range of images of asteroids with tail, I quickly generated a basic asteroid with tail image – a pixel wide asteroid at center with a horizontal tail with intensity going of as 1/r – and fed it to the pipeline. The two images on the right depict asteroids with tails. Not bad, eh?grouppresentation_mod-047

The image on the left is the only recorded image in literature of a trailed asteroid with a tail – 107P/1949 W1. This is an old image, on photographic plate.

On the right hand side I am showing you something similar that I tried to generate from my pipeline. The image on right hand side top is in a linear scale image while the one on the right bottom is the inverse color version, just like the 107P/1949 W1 image on left hand side. grouppresentation_mod-048

%d bloggers like this: