Very Faint Objects: Unselectable in PinPoint?

Some users new to PinPoint complain that PinPoint does not allow them to select objects that they can see in the image for measurement. There are two reasons for this behavior, both are covered here.

First Possibility: You Must Tune Your Settings!

If you have an obvious asteroid in the image, and you cannot manually click it and put a circle around it, chances are that your settings are a bit (or a lot) off. The size, brightness, sigma, and other settings that the user controls set the criteria that PinPoint uses to interpret the image. In effect, by adjusting these settings, you are allowing PinPoint to "see" stars and asteroids of various brightnesses.

If you get asteroids in your images that you can't put a ring around, but they sometimes measure a good, solid 3.5 above sigma (brighter than most of the stars visible in the image), then you can conclude you're running PinPoint with settings that are improper for your images. Check this out and tune it up right.

Second Possibility: Asteroid is Too Noisy

The second possibility is that the asteroid has no signal - or nearly none. In this type of case, the asteroid is visible on the image, perhaps with some difficulty. But statistically it is indistinguishable from noise.

Astrometric reduction software uses statistical methods to locate the center of an asteroid image (called the centroid). In short, it looks at the pixels that make up the asteroid image, and fits a smooth brightness profile to the pixels. Wherever this profile peaks is the center of the asteroid image, and that position is what is being measured.

Unfortunately, when an asteroid is visible on the image but does not have enough strength to rise above the noise floor, it can't be measured. If you did try to measure it, you'd be doing nothing more than measuring the position of a random brightness peak in the noise. Once you've measured a random position, do you really want to submit it?

The answer is no. Since asteroids typically take up three to nine pixels on an image, and a typical image scale is around 2 arcseconds per pixel, you could easily be reporting positions in error by as much as six arcseconds or more. The Minor Planet Center says that is unacceptable astrometry (if you were six arcseconds off, it would be unacceptable by a factor of six).

PinPoint differs from many, if not all, of the other commercial astrometric reduction tools in that it refuses to allow you to make this mistake. If the source that you are trying to click on does not have enough signal to be statistically significant, PinPoint won't let you put a ring around it and measure its position.

I consider this a competitive advantage:

• Stations that use PinPoint have smaller mean residuals than most stations using something else. Just have a look at the MPC's residuals calculations and see for yourself. Bottom line: if you want to turn in crappy astrometry, it is easier with some software package that will let you "measure" any old pixel in the background that you happen to think might be brighter than the others.
• People that want to measure ghostly asteroids that barely appear in the images, even when there isn't anything to reliably measure, aren't bringing a scientific attitude to their scientific work. Chances are, the work is seen as sportsmanship or discovery competition rather than as a scientific endeavour in these cases. PinPoint helps blunt this excessive passion, turns down the heat of the moment, and helps insure you aren't doing anything you'd be embarassed about later.
• It keeps people new to astrometric reduction from making a mistake during a time when they might not have the understanding to avoid it on their own.