Asteroid Zoo Talk

How often do you find asteroids?

  • lszatmary by lszatmary

    I made about 75 classifications, but I hadn't find any asteroids. Is it normal, or am I just too blind? 😃

    Posted

  • swbell by swbell

    It seems to be random. In my first couple of days I saw maybe ten asteroids, but nothing in the last three days!
    Actually there was one frame that the system said had a known asteroid but as far as I was concerned there was nothing there. So I might be just too blind too!!

    Posted

  • Cpt._Pete by Cpt._Pete

    I find at least 1 in every image..................But only 1 out of 20ish is real =)

    Posted

  • swbell by swbell

    It also depends on how realistic you are. A lot of frames have fuzzy transient spots which are obviously camera effects. It is easy to think these are asteroids as they may LOOK as if they are moving, but actually it is background noise. If the spot is very faint it has to move on every one of the four frames, before I flag it as a possible. It is highly unlikely there is an asteroid on every set you get.

    Posted

  • Dr.Asteroid by Dr.Asteroid scientist, admin

    Since this is real data, there are times the asteroids seem to be everywhere - and then there will be a slew of images where there don't seem to be any. It is a random distribution to be certain. Typically there should be about 1 for every other set - but that's a random distribution and sometimes you'll have more than one in a field and then a bunch without any.

    Posted

  • CTidwell3 by CTidwell3

    One every other set seems surprisingly high. I've reviewed over 300 images and found only 3 that I know of (plus one or two missed known asteroid). It makes sense what you are saying that there will be dry spells due to the random nature of which sets you get, but I know I commonly go through 10 image sets that appear to be good sets, spend a minute or two on each, and do not see anything other than stars.

    Posted

  • swbell by swbell

    I agree with CTidwell3. Every other set seems very high frequency, and often the quality of the image, or the amount of noise, would mean a very faint asteroid would be impossible to see.

    Posted

  • planetaryscience by planetaryscience

    Based on the fact that I've classified about 1,200 images, I would say that you get an asteroid approximately 1% of the time.

    Posted

  • TED91 by TED91 in response to planetaryscience's comment.

    1-2% sounds right

    Posted

  • Emili_Sancha by Emili_Sancha

    My percentage is 2.3% of known/unkown/untagged/suspicious asteroids

    😉

    Posted

  • dazmalski by dazmalski

    I'm interested in how many "unseen" images there are left? In the early days, obviously, they were very common. Now incredibly rare. Is there new data to sift through? Will we see new data interspersed within the data set or have we basically seen everything at least once?
    Ongoing new data or a just a set that has now been seen?

    Posted

  • theSkipper by theSkipper

    When I joined this project, I got several images with known asteroids, which I overlooked. Since then: no definite asteroid at all in over 100 datasets. I think I was being tested and, having failed, was given all the no-hoper datasets 😃

    I also notice that a lot of the images look r-e-e-markably similar to previous ones .. I see some others have posted that they are getting duplicate data.

    Posted

  • escholzia by escholzia

    IMO, the images are so poorly aligned and there is such a high % of bad images (missing frames, etc) and spurious pixels that it's surprising that anyone spots an asteroid without giving up first. I concur that the initial set was salted.

    Posted

  • rezoloot by rezoloot

    Theres nothing really to agree with, they seem to come in packs then nothing at all, the artifacts play a big part in any asteroid frame though, i don't trust any movement in artifact frames but as far as I know, the same has made me not circle two different asteroids that were accounted for, I guess we'll learn as we go along

    Posted

  • meegja by meegja

    I've done a lot of sets now and if I have to estimate is that around 3 to 4% at average are containing known and/or unknown [i]visible[/i] asteroids.

    There are rather a lot of non-vissible known asteroids: the system gives a green circle of a known asteroid but no way that it can be seen by human eyes. This either caused by the bad quality of the sets (very grainy/noisy, misaligned, bad set, etc.) or it is simply too faint to see but it is known by the system.

    So far the artifacts outnumber the asteroids a 100 to 1 I think. Or at least, it feels that way 😃

    Posted

  • theSkipper by theSkipper in response to CTidwell3's comment.

    Yes, I find the same - far fewer asteroids than DrAsteroid suggests.

    There are the original large images .. and then there are the small ones that we get to see. I wonder if DrA was thinking of the large ones, when he said "1 asteroid every other set".

    Posted

  • Meanjean4321 by Meanjean4321

    do you really have to see the asteroid move in all 4 frames. some that i find show up in 2 frames and i am pretty sure they are asteroids but i could be wrong. they move totally different than other items that are close

    Posted

  • CTidwell3 by CTidwell3

    You really need at least 3 for it to be a chance of an asteroid. Any two points can have a line drawn between them and appear to be movement in a direction. This is especially true when you know you have artifacts in images. Artifacts can appear to move in a straight line between any two frames, especially if you have the image on cycling and see frame 4 next to frame 1. It is only when you compare their movement between all 4 frames in order (1, 2, 3 then 4) that their movement becomes non-linear.

    With 3 points, you can confirm that all 3 are on a straight line. Still, because of the sometimes small amount of movement, having all 4 makes this determination much better.

    Similarly, unless the seeing is changing dramatically every 10 minutes, asteroids that are seen in one image should be able to be seen in the next image. Not seeing it, especially if you are dealing with what may appear to be a faint asteroid, may help determine that what is being seen is just an image noise artifact and not an asteroid.

    Take this image for example:
    http://talk.asteroidzoo.org/#/subjects/AAZ0001pc0

    If you look to the left just above center and look at just frame 1 and frame 4, someone might go 'Hey, ,there are two points there. that looks like an asteroid moving from right to left. It is just missing in frames 2 and 3'. Similarly, if you put the image on cycle someone might then go 'Hey, i was wrong, it appears to be moving left to right and just missing in 1 frame' (but incorrectly be seeing image 4, 1, (not seeing it in 2) then 3.

    The actually movement is where you see it in 1, up in 2, down and right to where it is in 3, then back left in 4 -- the typical movement of an artifact. Knowing this, you can sometimes eliminate moving spots seen in images by considering which frames you see things in and looking for things in the common locations. For example, see movement in frames 1 and 2 going upward... look in frame 3 down and to the right. See movement between 2 and 3 going down and to the right? Look in frame 1 below where you see something in frame 2. See movement in frames 1 and 4 -- look above where you see the object in frame 1. Etc.

    Posted