Asteroid Zoo Talk

asteroid, what to mark and not mark, i could be wasting my time

  • Meanjean4321 by Meanjean4321

    i see so many that could be asteroids but not sure. i would love to see vids of possible asteroids, what to mark and what not to mark. i might be marking a lot thats wasting my time and the peoples time that check them. i dont want to give them more work than they need

    Posted

  • Dr.Asteroid by Dr.Asteroid scientist, admin

    Let me see what I can do to post some unambiguous asteroids. Typically, if you're not pretty darn sure it's real, then don't mark it. There's no problem with that. Figure there will be about one asteroid every other frame (although sometimes there are more than one per frame).

    Posted

  • theSkipper by theSkipper

    Dr Asteroid, your estimate of one asteroid per 'frame': do you mean one per set of four images that are presented to us? Because most people seem to be reporting at most one every 100 sets of images!

    Or do you mean one asteroid per Catalina Schmidt image? This would be about right, as each 4096x4096 image is split into 256 smaller ones which are presented to us (?).

    Posted

  • Dr.Asteroid by Dr.Asteroid scientist, admin

    Ah! To clarify - we would expect roughly one asteroid per set of four images. (256 x 256) We settled on that size as a good estimate of what a person would be able to keep an eye on - and have typically something in there.

    But it's random - and not all the images will have one. Some will have more. There are hundreds per large Catalina Image.

    Posted

  • CTidwell3 by CTidwell3

    Currently I've evaluated 2035 sets and have 60 flagged as seeing an asteroid where what I thought I saw was not a known asteroid. I don't have a count of how many known example sets I've come across, but if I take a range of double (60+60) or even triple the count I flagged as new asteroid (60+120), we are talking 1 asteroid every 12-17 sets reviewed.

    Add to that the mis-aligned, bad_sets and star-bleed/over-exposed sets. I reviewed my last 120 sets and visually counted 20 of these were mis-aligned or star-bleed. Estimate another 10 bad_set records (only 1 or 2 images type of thing) and that is about 25% of the record sets that can't be easily evaluated. If I then modify my evaluated count using this, it would be more like 1526 sets evaluated -- meaning one asteroid every 8-12 sets. Not as bad as the 1 per 100 that it feels like at times, but still not the one every other set rate that is expected.

    I know there is a random nature in this, but to have one every other set would imply that for every time I go many sets without seeing an asteroid, there should be at least some times where I see asteroids in consecutive sets. I don't remember ever having that happen to me.

    Now, given that there are hundreds seen in a typical large Catalina image, I'm guessing that the discrepancy between expected sets viewed per asteroid spotted and what is being experienced may be revealing some interesting data. For example, this data could be showing that the details we can see in these 256 x 256 sets and/or their alignment using the tool here is making it harder for some fainter asteroids to be noticed as compared to viewing the original Catalina images through other means. It's all good data to collect, since it will help fine-tune what we do here as we continue reviewing data these data sets.

    Posted

  • Puppyhogg by Puppyhogg

    what is the meaning of "keep an eye on"

    Posted

  • CTidwell3 by CTidwell3

    In this case it would mean being able to monitor changes without becoming too overwhelmed with data. If too large of an image size were selected, there could be so many stars in that one image that some would give up and say there is too much to look at. I could become easy to forget if one had checked each point or not.

    Posted

  • Dr.Asteroid by Dr.Asteroid scientist, admin

    We did generate a series of 512 x 512 images and everyone involved agreed those were too big - there just was too much going on to keep track for anyone to work with it.

    Posted