A Method for Eliminating or Qualifying Known Objects Using Background Stars and Other Objects

reply to post by dainoyfb


Slightly off topic, but I appreciate what you're trying to do. Arguing over this sort of stuff is essential and we all benefit from it.

One thing that I find interesting regarding getting a qualitative idea about a light in the night sky is amateur spectroscopy:

www.amateurspectroscopy.com...

Can't pretend it's a new idea, looks like amateur astronomers have been doing it for some time, and I heard of it from one of the Hessdalen related papers:

www.abovetopsecret.com...

The quote from the paper is that a basic spectroscopy grill costs about $200. But what I like about the idea is that because of the interest of amateur astronomers it seems to be fairly well documented.

reply to post by dainoyfb

Good thinking, with a few caveats. You don't know the actual distance between objects flying in formation, only that it is greater than the supposed wingspan, which is already useful.

A second remark, this technique has been used (by me and others) in a thread about a year ago where a long exposure photo showed two trails of lights. From the typical wingspan and speed of an airliner, smaller plane, model it was possible to check that it couldn't be any of them. Why ? Perspective. Angular distance and exposure time do not translate to speed when the size (wingspan) is known, because there is a radial component that is not seen.

[edit on 2010-1-5 by nablator]

Originally posted by dainoyfb
We will first test for the possibility that the objects are Canada geese flying in formation. We know that Canada geese have a wingspan of approximately 2 meters. The objects according to our star field reference are about 0.38 of a degree across.

I like your idea and math is my strong suit so I think your math makes sense. But where the idea might fail is assuming the photographic image has any relationship to the size of the object being photographed, especially when the object(s) photographed are fuzzy dots. I'll give you an extreme example here:

[atsimg]http://files.abovetopsecret.com/images/member/28dad4b6e880.jpg[/atsimg]

That's a screen capture from the famous STS-75 tether video and I think the width of the tether (below the yellow arrow pointing to a star) is probably about 3 orders of magnitude greater in that image than the actual object width. The tether is about 2.5mm wide and is photographed from a distance of about 75-100 miles away so it should be so thin as to be invisible at that distance if your method worked. The difference between geese and planes in your example is only about one order of magnitude so you simply can't trust the camera to capture an image size that relates to the size of the object if the object is relatively small. Or to put it another way, photographic conditions would permit Geese showing up on an image as larger than planes under certain conditions when of course the geese are much smaller. I think this is particularly true of say, white birds, which will reflect more light relative to their size due to the high reflectivity of the color white, when compared to a darker colored plane, like the F117 or even a plane not quite that dark.

Now here's where your method WOULD work I think:

If the object photographed is large enough to be something besides a fuzzy dot (or fuzzy line in the case of that tether), I think it would work. I saw you dismissed pixels earlier and claimed your method has nothing to do with pixels and I understand your objection, however I think if it's a digital image you could say the more pixels the image has that you're trying to estimate the size of the better your method will work. Less than 10 pixels I think the errors will be large, but if the object is say 30-50 pixels at a minimum in size then perhaps your method has some merit.

Star and flag for coming up with a good idea that will work sometimes, but only with "larger" objects or more specifically objects with a larger angular size in the image (regardless of the actual size of the object).

Unfortunately so many videos are of fuzzy dots I would almost say the majority seem to fall into that category and that's where the camera limitations and not the math prevent your method from working.

[edit on 5-1-2010 by Arbitrageur]

reply to post by Arbitrageur


Yes, I agree that image quality will have a lot to do with the accuracy. Hopefully stacking and other types of processing will help in some cases.