The purpose of this document is to define standard procedures for examining Mosaic frames for the Deep Ecliptic Survey with "LOOKER" and reporting results (to Marc Buie) on personal LOOKER log sheets.
The objective of the frame examination is to mark KBOs and other moving objects (Centaurs, comets, Trojans, and NEOs) in a manner to facilitate the next step in the process: the astrometric analysis for follow-up observations. Hence the frame examination should avoid the following errors: (i) missing any of these bodies and (ii) reporting highly questionable candidates. The former error reduces the efficiency of the survey by not detecting KBOs that are in the data, while the latter error reduces the efficiency of the survey by wasting time at the next step of the analysis.
Our current procedure is to record two frames of the same field with Mosaic, separated in time by about two hours. The frames are corrected for bias and flat field. Pairs of frames are displayed by "LOOKER", one in red and the other in cyan. Hence a moving body will appear as a red-cyan pair, which can be marked with "LOOKER" as a moving object. In the subsequent stage of the analysis, objects marked with LOOKER will be examined in two ways: (i) a Viasala orbit will be fit to the two points (objects with unusual orbits will be discarded), and (ii) the "devil's advocate" will re-examine the images (objects appearing to be cosmic rays and other artifacts will be discarded). The devil's advocate likes KBOs but hates false alarms.
When marking images, the left mouse button is used to measure the moving object on the "red" or primary image in the main image display area. Measure the "cyan" or secondary image using the middle mouse button in the main image display area. The right button is used to select a different moving object that has already been measured by clicking near its position in the primary frame. When you click on an object in the main image display area, the automatic centroiding algorithm is always used. You don't need to click very accurately and you cannot affect the final psotion in this window. However, it quite often happens that the automatic centroiding does not lead to a good measurement of the position.
The zoom window lets you see how the synthetic photometric aperture is located relative to the object. If the circle seems poorly placed you can move it by clicking in the zoom window. The left button will reposition the circle without changing the location of the zoom window (and the stretch). The middle button will move the circle and recenter the zoom window and will reset the stretch. On any slowly moving objects you should always be looking at the position as shown in the zoom window. If it doesn't look good, fix it. In principle, you should do this for all objects but this can be a lot of work for the main belt asteroids and is optional unless the position is really, really bad.
Currently LOOKER is set up to be run in the following computer environments: (i) the computing network at Lowell, (ii) computers at MIT. At Lowell, the computers are set up to read and write files from a single location: /gryll/data6/buie/reduced/yymmdd, where yymmdd (yy are the last two digits of the year, mm the two-digit month, and dd the two-digit day) is the directory containing the reduction files for a particular night's data. Using LOOKER from the Mac laptops or from astron.mit.edu requires porting files from the gryll.lowell.edu location, writing to them (LOOKER does this), and then porting the files back to gryll.lowell.edu.
One should first get her/his LOOKER assignment from the web page appropriate to the data set being analyzed. The URL is addressed according to the date http://www.lowell.edu/users/buie/kbo/lookyymmdd.html. This page should be printed, and it becomes the LOOKER log sheet. The name of the examiner should be written at the top of each sheet in the space provided. In some cases the assigned frames will have been already examined by the computer or one or more individuals. If the frames have already been examined, then the results of the examination(s) will be stored in "obj files" on gryll.lowell.edu. In not, then LOOKER will be creating the obj files as the frames are examined for the first time.
When working in the Lowell environment, the Mosaic frames, obj files (if they already exist), and the script ("looker.pro") for invoking LOOKER should be in place and require no special action on the part of the user. Working outside the Lowell environment, on the other hand, will require the user to be sure that the Mosaic frames, obj files, and the script file for invoking LOOKER are all accessible and have been set up properly. The Mosaic frames are in FITS (binary) format, and the other two file types are in ASCII format. Hence when porting these files from gyrll.lowell.edu, one must be sure that the ftp program is set to binary mode for the Mosaic files and text (ASCII) mode for the other files.
The mechanics of using the LOOKER program are given at the URL, xxx. One should learn how to display an existing field for examination, create a new field by opening two corresponding Mosaic frames, align two frames, mark (and unmark) a moving object, and classical frame blinking.
If the field has been previously examined, the two frames will likely be aligned so that the stars and other objects appear white. On the other hand, the two frames need alignment (i) if the LOOKER display appears as collection of red and cyan objects, or (ii) if the objects are red on one side and cyan on the other. When aligning frames, the best results will accrue from a star that is not saturated and near the center of the frame. When the two frames contain a relative distortion, due to refraction in the ADC or the Earth's atmosphere, two or more different alignments may be needed to display the objects as white on all parts of a single frame.
The normal procedure is to examine the frame pairs and match up red-cyan image pairs. Normally the red and cyan images will be of similar brightness and, if the body is moving fast enough, the images will be elongated (due to the motion) along a line between the two images. Sometimes one image is clearly visible, but the other is hidden in the wings of a bright object. In some of these cases, the second object can be seen. Not all red and cyan objects are moving objects. For example, a cosmic ray strike will appear on one frame but not another. Other artifacts occur, and one must differentiate these from real moving objects.
The types of moving objects that we encounter in the Mosaic frames are given below. The (retrograde) rates refer to bodies at opposition. The rates decrease further from opposition until they become prograde. Note that you should always see an amount of object trailing consistent with its motion.
| Object | Characteristics |
| KBO | stellar appearance, moving less than 5 arcsec/hour retrograde |
| Centaur | stellar appearance, moving at 5-15 arcsec/hour retrograde |
| Trojan asteroid | slightly elongated, moving about 20 arcsec/hour retrograde |
| Comet | fuzzy image, can be moving at any rate and direction |
| NEO | very elongated, moving 75 arcsec/hour or faster, can be moving in an odd direction |
| main-belt asteroid | elongated images, usually moving 30-40 arcsec/hour, retrograde |
As you might guess, KBOs are the central purpose of this project. That's not to say that the other objects aren't interesting but it does reflect the priorities for processing and followup that we must set. A KBO is relatively easy to followup. We have been successful in predicting and confirming KBOs as much as 3 months past the time of discovery. This long window for recovery is due to two effects: (1) these objects move slow and can't go far, and (2) a good expectation on the orbit for the object.
All other objects are increasingly difficult to followup in that they require further observation much quicker than KBOs. We have roughly 1-2 weeks to catch a Centaur before it will be lost. These objects don't move too fast but their orbits can be just about anything making the search volume much larger. Trojan asteroids are probably recoverable for a month. Near-earth objects (NEOs) are very difficult and must be recovered within at least 24 hours. Main-belt asteroids are probably good for a week. Comets can do just about anything, for these the sooner the recovery, the better.
Our data processing and examination pipeline usually precludes reliable followup of everything except KBOs though Centaurs are feasible at times. All other categories of objects are beyond the scope of the project and will generally not be followed up. If you, as discoverer of a non-KBO, want to see a specific object followed, it will be up to you to coordinate and obtain followup observations. All observations of all objects will eventually be reported to the Minor Planet Center on behalf of the team. Followup and continued work by the team will only take place on KBOs though Centaurs will often be tracked as well since there will be far fewer of them.
| Object Type | Marking Style |
| KBO: | xn:mm |
| Centaur | xn:mm Cen |
| Comet | xn:mm Comet |
| NEA | xn:mm NEA |
| Trojan | xn:mm Trojan |
A question mark flags an object of any type you believe is of dubious reality but you can't fully decide to ignore (example: x8:04? Cen for a questionable object that could be a Centaur). This marking should be used very sparingly. As a rule of thumb, less than 10% of the KBO and ? objects should be ?. If you only mark bonafide 10 KBOs, then you'd probably be justified in marking no more than 1 ?.
Centaur's should be treated the same as described above for questionable objects. Statistically speaking, Centaurs are much rarer than KBOs. If you are marking a lot of Centaurs then you probably aren't being critical enough in your image examinations. Comets will be even more rare. We have no real idea of the statistics involved but we might be surprised to find more than 1 comet every three years of observing. If you are marking 5 comets on one night you are not being discriminating enough.
On the first pass you are looking for new objects on a blank slate. Measure any and all moving objects as you go. On the second pass you are looking for stuff missed in the second pass. Unless the first pass is very, very obviously in error, don't modify or otherwise delete any first pass objects. Object verification and weeding will happen after all the passes are done by the aforementioned "devil's advocate".
When an assignment is completed, the LOOKER log sheet is FAXed to Marc Buie (520-774-6296). The LOOKER script and the obj files are ported to /gryll/data6/buie/reduced/yymmdd. Be sure that the permissions on the files after they are ported to gryll are 664 (readable and writable by the owner and group; readable by all). Note: Do NOT resubmit log sheet information. Once you have FAXed or otherwise turned in the sheet you must print a new copy of the log sheet before making additional marks. It is better to work from a version of this sheet that has been updated but it is not required. Bottom line, every mark on a sheet turned in for entry should be a new mark that needs to get entered.