(25143) Itokawa


Asteroid Hunting




On November 17, 2009 I found my first real asteroid.  While it was not completely accidentally, it certainly involved much luck.  My big rock was originally discovered in 1986 and goes by the handle “4035 1986 WD”.  I found it while taking 5 hours of images of M74.  The “Hunting” part of Asteroid Hunting does not really apply to me at this point.  While I was looking for a comet, asteroid, or minor planet in my images that day, I had not pointed the telescope in any particular direction, nor had I captured images from the CCD for the purpose of finding these transient objects.  I was happy trying to take better and prettier pictures of the M74 and other deep sky objects.

Each year I have been attending the Advanced Imaging Conference (AIC2009).  Over the last couple of years some of the professional astronomers have mentioned during their presentations that we amateurs should not be too quick to run our statistical data correction routines to clean up bad pixels in our images.  The reason is that while these advanced procedures will make our images cleaner and more appealing; they also remove from the images any small aberrations including asteroids and minor planets.  Normally, during image processing, I will run a Poisson Sigma Reject algorithm over a stack of aligned subframe images of the same object to remove hot and cold pixels, satellite trails, meteor trails, and cosmic ray hits on the CCD.  These adverse blemishes are plentiful.  The professionals were telling us to take the time to look at the raw images first, before running our statistical methods. 

While I had done this a couple of times before, I had never seen anything that looked like an asteroid to me.


What really makes this hobby fun is discovery, doing something for the first time where you were never sure that you had the capability to do it.  Finding my first galaxy back in 2005 was such a moment.  (Finding M51).  While my asteroid is a big rock, and maybe even classified as a minor planet, it is still very small on the scale of our solar system.  Its orbit is at 5.04 AU’s from the sun, which means that it is very close to Jupiter, which is at 5.2 AU’s. [1AU (Astronomical Unit) is defined to be the distance from the Earth to the Sun].  I was not expecting that I could find something that small at that great of a distance from Earth.

 On November 19th, the day I was processing the images captured on the night of the 17th, I remembered to look before doing a data cleanup of my stack of images. Here is what the inverted image would looked like if I run the Poisson Sigma Reject algorithm with σ = 1.6, and here is what the raw inverted image looked like.   I have pointed to the asteroid in the raw data image, and you can see that it is gone in the cleaned up image.   So the first step in discovery is to not throw away what you are looking for.

Blinking of stacked images and observing any object movement is the classical method of discovery.  This has been used by professionals for decades to find things in our solar system that have movement faster than the sidereal rate (how fast the stars appear to move due to the rotation of the Earth).  Like building a cartoon with animation, if you blink or quickly page between images the objects can be seen to move.  This movement points to things within our solar system that are not background stars, but something that has a much faster orbit.  Some of the more distant planets were originally discovered with this method.

The problem for me anyway, is that you need a lot of patience and careful observation to do this.  I can well imagine that this process has already been computerized.  But I have not found a program that does it.  So for me it is a slow manual inspection, and this is painful to do.  I do have software that blinks the stack, but you still have to look carefully to see anything.  Before the data from November 17th, I had never seen anything.

I use CCD Stack software for initial image processing.  During the stack registration process which aligns all the images in the stack, I observed a small black spot among all of the white stars on a couple of the raw (pre-Poisson) images.  At this stage CCD Stack provides some translucence of the top image over the base frame image of the stack, and the color black is used to highlight the absence of any pixel match with the base frame image.  This was unusual enough for me to spend a few more moments investigating. Here is what triggered my interest to look further, just this small dot.

Rather than using blinking, which to me seems very easy to overlook something, I used a technique that I again learned at the AIC2009.  I simply summed the already aligned images into one image.  This produced this revealing image, which I have inverted to make it easier to see.  The presence of a faint line in the image created by summing the stack of 20 images was a direct observation of a moving object (just like a simple cartoon animation).  At this point I blinked the stack, now knowing where to look, and had discovered my first asteroid. Click here to see an animation of the movement of the asteroid from images that spanned about 1.5 hours of exposure:


Now the obvious question… what is it?  Is it something never seen before, or is it a previously discovered object?  In my case, I knew that the size of the object seemed to be pretty large, so I had no eager anticipation that I had found something new.  In fact since it was my first such discovery, I really needed to confirm what it was, and if it was a new discovery then I would not be able to confirm anything.

I had been reading about various MP Checkers which are made available from Harvard University.  Here is a link if you are interested: http://www.cfa.harvard.edu/iau/mpc.html  These internet tools will give you the ephemeris data for orbital objects if you provide the precise time and celestial coordinates of the object.  In this way, you can determine if the object has previously been discovered.  Therefore once you have the Universal date and time (which is captured automatically along with each CCD image and is in the .FIT header of each file) only the precise celestial coordinates of the object still need to be determined.  Even before you know the coordinates, The Sky6 Planetarium software also can display hundreds of thousands of asteroids.  Actually I found my rock first in Sky6 then confirmed it using MPChecker.

The coordinates are not that difficult and I have been able to do this part for several years now.  I first discovered how to do this when trying to find Pluto.  There are two software products that I use to perform a “Plate Solve” which is now an old term for a computerized process that aligns the image with its Planetarium position in the known sky at a specific date, time and location of the observatory.  I seem to have the best luck with this using MaximDL software, but Sky6 along with CCDSoft will also do the same thing.  Once the mapping has been successful, moving the mouse cursor over the CCD image will display the RA and DEC coordinates at that point.  I found my moving object to be at (using J2000 coordinates) RA 01 Hrs 36 Min 44.47 Sec and DEC 15 39’ 19.8” at UT 05:39 on November 18, 2009.  Which is “2009 11 18.23” when entered into MPChecker.  Dates and time changes are a pain. In New Mexico at the time of the image it was 10:39pm MST on November 17th. 


Here is the report that you should receive back from MPChecker.

The first line is my Asteroid (4035) 1986 WD.

Notice that the coordinates are not exactly perfect.  The RA is off by .33 seconds and the DEC is off by 3.2 seconds.  I am not sure exactly why the difference other than I used observatory code = 500 which is the default entry and I have no idea of its real location relative to the location of my telescope near May Hill, New Mexico.  This has to be part of the error.

Moving from pretty pictures to the world of real science:

Now the question for me is how far and how fast do I want to move into making real scientific contributions.  Every day there are requests for armature astronomers to contribute location data of specific Near Earth Orbiting Objects (NEOs).  These data are used to help model and determine accurate orbital formulae for more newly discovered NEOs.  This is a real contribution and critical in projecting future object locations and possible impacts.

To get started with this I would need to establish my own Observatory code with NASA.  There are preformatted computer interface records prescribed for both sending input and receiving feedback.  The forms have obviously all been computerized by the big national observatories. 

Am I ready to do this? Actually trying to target asteroids and minor planets would radically change my imaging targets. Target strategy would be interesting. The location of the Asteroid Belt is well known and it holds millions of these objects. I know one amateur astronomer who prefers to search elsewhere thinking that the big university observatories with their fully automated sky searching computer systems will find most objects in and around the belt. He looks were fewer people tend to search. Because these object move so quickly, even if someone observes a pin-point location in the sky one night it does not preclude the same location having an object just a few days later. Three days after I discovered 1986 WD I tried to take more images of it in the same location, but it was no where to be found.

I could actually become an asteroid hunter?

 November 20, 2009

November 22nd ... I was successful last night in re-acquiring 1986 WD, now at a much different location. To prove to myself that I can find Asteroids I also targed a new object 2001 YZ124. Can you see it?