A guide to artifacts
-
I thought this might be helpful for anyone looking for what to count as an artifact and what not to. The official page on it is helpful, but incomplete so I'll put this here as a more complete guide, and if anyone notices a new bug feel free to post it! 😉
Artifact: Cosmic rays
Appearance:
white dot slightly above center
AAZ0000zzp
cause: cosmic rays
Artifact: "hot" pixel
Appearance:
white streak below and left of center
AAZ0001s8k
Cause: a broken pixel not collecting light properly
Artifact: dot with hat
Appearance:
top right quadrant
AAZ0001sjb
cause: unknown
artifact: star bleed
Appearance:
also 
AAZ0000n7r
cause: a bright star
artifact: bad image
Appearance:
AAZ0000jo4
cause: unknown
Artifact: FMO
Appearance:
AAZ0000yvg
cause: a satellite or artifact.
Artifact: Lens flare
Appearance:
AAZ000170g
cause: a ghostly reflection of a nearby star on the telescope
Artifact: Dark star center
Appearance:
AAZ0000hj5
cause: the star being too bright for the pixels to handle, so they black out.
Artifact: 'flat field defects'
Appearance:
left of center
AAZ00019rh
cause: unknown
Artifact: Comet (not really an artifact)
Appearance:
credit: @scibuff
AAZ0000qwx
cause: comet
Artifact: 'blinking' satellite trail
Appearance:
credit: @scibuff
AAZ000049p
cause: spinning/blinking satellite
Artifact: haze
Appearance:
AAZ0001ugh
cause: haze high in the atmosphere interfering with observations
Artifact: 'bad pixel coloumns'
Appearance:
AAZ0000q8c
cause: unknown, possibly similar to hot pixel artifact.
Artifact: meteor trail
Appearance:
AAZ0000n36
cause: a meteor from a meteor shower
Artifact: FMO
See Satellite trail/Blinking satellite trail/meteor trail above.
Artifact: ?
Appearance:
AAZ000063w
cause: ?
Artifact: Diffraction spike
Appearance:
AAZ00011tm
cause: a bright star
post any others you know of below!
Posted
-
by
Staceva
Very helpful thank you 😃
Posted
-
by
enderb
This explains a lot. thank you so much!!
Posted
-
by
Cpt._Pete
I just say to myself....
There are 3 things in the pictures,
1,Stars
2,Stars that move (Asteroids)
3,all else is Artifacts=)
Posted
-
by
scibuff
Hmm ...
http://talk.asteroidzoo.org/#/subjects/AAZ0001dn8
does not seem like a cosmic ray, you can see that pixel move from frame to frame in the exact same pattern as other artifacts. I'd guess it is a bad CCD pixel that moves around relative to the background due to dithering.
update: this seems much more like a cosmic ray >> http://talk.asteroidzoo.org/#/subjects/AAZ0000zzp
Posted
-
by
planetaryscience
in response to scibuff's comment.
Thanks for that. Updated along with a few more artifacts added
Posted
-
by
scibuff
in response to planetaryscience's comment.
any chance for "sticky" posts? this def deserves that status
Posted
-
by
scibuff
in response to planetaryscience's comment.
Comet - http://talk.asteroidzoo.org/#/subjects/AAZ0000qwx
not a comet - http://talk.asteroidzoo.org/#/subjects/AAZ000049p
Posted
-
by
planetaryscience
in response to scibuff's comment.
I would if I could, but only moderators can 'sticky' posts. 😦
Posted
-
by
scibuff
in response to planetaryscience's comment.
I thought you were one (looking at the username 😄)
Posted
-
by
planetaryscience
in response to scibuff's comment.
No, unfortunately I'm just another contributor, but I still try to help as much as I can around here and elsewhere. 😉
Posted
-
added this artifact
and updated data on FMOs
Posted
-
by
waylandwx
Interesting posts.
Posted
-
by
jsalvata
How about those large dark dots? I've also seen white circles...
Posted
-
by
planetaryscience
in response to jsalvata's comment.
The white circles, I'm assuming, are similar in nature to the dark circles. They are #8 on the list, I believe, and any ideas/suggestions you have for what causes them would be gladly accepted. I'm stumped.
Posted
-
by
scibuff
Here is what a fast moving object (FMO) looks like
http://talk.asteroidzoo.org/#/subjects/AAZ0000e1u
Posted
-
It appears that FMO has been assigned as a name to anything that moves fast.
Posted
-
I'll just bump this up again for the easy accessibility of readers.
Posted
-
by
Emili_Sancha
a blinking satellite...
Posted
-
Yeah, sorry I've been inactive lately.
Posted
-
by
lszatmary
Where should I put my artifact mark on a long vertical pixel string or on a long diffraction spike? At its middle maybe?
Posted
-
I don't suppose it matters, if you mark it I think they'll see what the object was.
Posted
-
by
lszatmary
OK, thanks for the answer.
Posted
-
by
Emili_Sancha
I suppose that diffraction spikes and vertical saturation lines are "common" in camera artifacts. The problem, i think, will be the ramdom dots that appear in the images.
Posted
-
those certainly are a nuisance, but a to-be-expected side effect of 'noise' in the images.
Posted
-
edit: updated data on star bleed and diffraction spikes
Posted
-
by
CTidwell3
A small update for these two:
Artifact: black circles and Artifact: bright pixel string
Taken from the spotters guide: "Other types of artifacts may be visible in the images. Flat Field Defects (A) or Bad Pixel Columns (B) will appear as shown in the image above."
Doesn't give a reason for them, but is at least an 'official' name for both of those artifacts.
Thanks for the work you have put into this post -- it is very helpful.
Posted
-
updated accordingly.
Posted
-
bumping.
Posted
-
by
Dr.Asteroid
scientist, admin
Should be locked to the top of the discussion now.
Posted
-
by
bc2callhome
How long will it be before we get any results and can we name any new asteroids
Posted
-
It will probably be a while, after the first set is all classified. Since the asteroids are found by a number of users, I would assume that normal rules for the naming of asteroids would need to be changed.
Posted
-
by
stonepenny
Thank you - very helpful - I've had some of these, and wondered. Would you please include planets? Had one set, marked 4 dots going around a star! I marked as 'asteroid' at the time... 😃
Posted
-
by
stonepenny
#asteroid or planet? Above biggest middle star, going from right to left, tracked over 4 marks
Posted
-
Exolanets emit light only in the infrared spectrum, and are much to small and close to their star to be visible from these images*. As a result, the 'planets' you see are artifacts and asteroids.
*I actually did a study of exoplanets that orbit their star distant enough to see with this size of telescope, but like I said they're extremely small. Like somebody said once, it's like looking at the head lights of a car in New York from Los Angeles- except this telescope can only make out a large building.
Posted
-
by
rnairn
Out of curiosity - what is the artifact data set being used for? I wonder how much time I should be spending on labelling artifacts. Labelling multiple artifacts is a bit of a pain. It would be nice to have a "label artifacts mode" where you can turn it on bulk label. Handy for when you have the blotches on multiple frame sets. In addition, perhaps Bad Set could be an additional button which would flag the set as unusable for classifying rather than treating a blank or corrupt frame as an artifact... Other than that I enjoy the project.
Posted
-
by
2135236
Hello, I am just wondering what is zooniverse for?
Posted
-
by
CTidwell3
@rnairn - Part of the answer to that question can be found here:
http://talk.asteroidzoo.org/#/boards/BAZ0000001/discussions/DAZ00002xf@2135236 - Zooniverse's About Us - Purpose page has the best answer to that:
https://www.zooniverse.org/aboutPosted
-
by
alpha61
A question on artifacts; do I mark the artifact on the first image then move on, or do I have to mark the artifact one each of the 4 pictures? Thank you.
Posted
-
by
CTidwell3
@alpha61 - Ideally you would mark them on all 4, assuming they show up in all 4. This type of data, when used, will likely be for enhancing future computer algorithms for identifying artifacts, so that would need them in all frames.
Posted
-
by
Antigonet
If there was a way to review other people's opinions of sets I have done (anon would be fine), I think that would improve my accuracy. Someone mentioned badges..... knowing how many studies a reviewer has seen, coupled with knowing what they thought, might be useful.
Posted
-
by
AstroTinker
If you were the first to see them, and you commented on, hashtagged, or collected them, you might pull it back up after a while and see if anyone else who has seen them has commented. Some will comment just in response to your comment or question. If you see others marked similar, you have that feedback. If they don't hashtag or comment, you won't know if anyone else has seen them yet.
Posted
-
by
nicro46
If I can give my opinion, what is lacking in this program is the support of one or more moderators (as in other programs Zooniverse) who can comment in real-time (or almost) our work. This involves, especially for newcomers, the images often bring into question not significant, that have nothing to do with the scope of work. it is obvious that everyone must have experience, and in this case can help a moderator to confirm whether or not a result, and maybe to suggest that do not bring in Talk too many images of artifacts , cosmic rays, FMO and many disorders that are high in almost each set, in order not to have too many images to comment
Posted
-
by
DZM
admin
in response to nicro46's comment.
Moderators are appointed by the research/science teams, so if the team would like to add a moderator, I'd be more than happy to do so...
Posted
-
by
nicro46
And we all appreciate it very much, DZM .....
Posted
-
by
AstroTinker
in response to planetaryscience's comment.
Some observations from my couple of months here checking images, reading posts, and researching online...... There are several classes of artifacts which seem to move between images in a particular non-linear pattern or two, sometimes in groups like moving 'in formation'. The basic reason for the apparent motion relative to the background starfield is telescope repositioning. The telescopes for our surveys take a set of four series of shots, each shot lasting about 30 seconds, and each series finishing in about 10 minutes. When it finishes each series, it tracks back fairly closely to where it started the first series, and does the next series. Thus, our image sets are, for example, shot 'h' from series 1, 2 ,3, and 4. Since the telescope repositions (and the earth moves) between each series, the exact center of the telescope lens, and the imaging CCD camera, is not exactly in line on each image of our sets. Thus, when the software tries to align the starfield in all four images, any defect anywhere in the light path of the telescope would appear to 'move' in relationship to the background starfield. The pattern of movement may be specific to a particular telescope, or maybe even to the general portion of the sky. I have noticed two primary patterns of motion of these defects, one which resembles an upside-down(inverted) T, and the other an upside-down Y. Thus, for the inverted T, image 1 is the baseline, and the defect would be at the junction of the T. In image 2, the defect appears to 'move' up relative to the starfield; in 3, back down almost to the level of 1, but to the right; in image 4, the defect appears to move to the left, about as far to the left of 1 as 3 is to the right. This means 1,3,4 appear nearly in a row, but in sequence 4,1,3. The inverted Y is similar, except 1 appears to be near the center of all 3 others(no 3 in a line). [Sorry, but I cannot illustrate this in text in the boards]. ----The main take-away from this is that artifacts that appear to move in this fashion are NOT real object(s) in the starfield, but defects in the light path of the telescope.---- ....Next post... some -possible- causes of the defects themselves.
Posted
-
by
Cdw24234
Hello, quick question - hopefully not a repeat question, apologies if so, have only been active on the site for the past week or so. I've noticed that sometimes in my image set I'll see the appearance of what looks to be a star (for lack of better comparison), but it will only be in one of the frames whereas other stars stay steady. For example, I might notice it on set 3 of the series, but it will "disappear" for the other sets. I am marking these as artifacts, but I'm not sure I should be as they really don't appear to be a "hot pixel" or moving satellite, etc. Any thoughts or recommendations on if I should continue to mark these?
Posted
-
by
AstroTinker
in response to Cdw24234's comment.
Cdw, sounds like you are seeing flat-field defects, basically dust or something in the light path in the telescope. Look for explanations of the motions under the tag #ffd (I have several notes on them)
Posted
-
by
MvGulik
Diffraction spike:
Can someone confer that only having one* diffraction spike is possible / normal ?
*) the diagonal one. (as the vertical one's are actual diffraction spike too if I'm correct.)
Source star: " * alf Boo" -- Red Giant Branch star.
Source CSS: 01_12APR16_FD1608.(darn ... Line 2 should be italic too.)
Posted
-
by
MvGulik
Just dumping something I'm still pondering about.
Identifying FMO's.
-
Generally kinda hard to not possible in a conclusive way.
-
Short one's with a length of about 75% width/height of Azoo/CSS image: High probably its a Geo-satellite. Generally having the same streak brightness (with or without flashes).
-
Longer ones, can be Lower Earth orbit satellite, or long streak meteor.
-
Meteors are tricky one's. as they can range from a small spot (perpendicular motion to telescope) to a long streak covering the entire master-frame. (general height 100 to 70 km) ... best identification feature: clear smooth change in brightness, preferably with some irregular flareup. (generally coming from the area the Earth is moving towards ... which is a general unknown with a lot of the CSS images.)
-
Airplanes, seem to have generally master-frame size streaks, and generally having a fussy nature. due to there relative short distance to the telescope (ergo: a bit out of focus). But so far I have not seen any track that I could clearly identify as coming from a airplane. Plenty of faint and fussy tracks, but no additional useful track features spotted yes.
Posted
-
-
by
MvGulik
A little bad image update.
Artifact: bad image.
Appearance:
AAZ0000jo4
Cause: A nonsensical decision by the (initial) Azoo project team.Let me clarify this a bit.
That seemingly junk column of data on the left side of this image is no real junk data. Its additional data that is pulled, together with the master-frame, from the CCD camera system. And one main purpose of this data is to be able to detect, and therefor also correct, any CCD introduced imaging bias. (search for CCD + overscan and/or prescan for additional details, details which are not relevant to this post)
Needles to say that this additional data is only useful when doing master-frame image corrections, and that it has absolute no usefulness for the Azoo users. And even worse is that the inclusion of this additional CCD data is the reason all Azoo sets that are coming from the first grid column (grid-cells 1..17) of a master-frame are the way they are.
Additional ...
The fact that this overscan data was included in the Azoo-set images, and the fact that grid-cell 1 is located in the absolute bottom-left of the master-frame. Strongly suggest the Azoo team did not much thinking and testing when they decided how to chop-up the the supplied CSS master-frames.By going for the absolute bottom-left in the master-frames they also made sure that a lot of master-frame bottom-sets (1,18,..,273) ended up with clipped images. (Which is due to the small telescope offset's between the different time-frames.) + (this would not have been such a annoying issue, if the classifying page would not auto-center the image.)
Example: Set AAZ00005mi, which is a grid-cell 1 set.
I'm no professional scientist, nor a professional coder. But I do know when general common sense was overruled by something else. Which seems to have been the case here.
[Bad Column-1 Image-set](#/boards/BAZ0000001/discussions/DAZ00000bi?page=6&comment_id=56890d2ca243c3496100004f)Posted
-
by
Canmore1919
So are we supposed to note all of these types of observed features as artifacts whenever they are seen on an image? For instance I have seen so many star bleeds that I am now ignoring them.
Posted
-
by
MvGulik
Nope. You can safely ignore them whiteout giving it a second though.
As far as I can see there seems to be no good reason's for tagging any artifact, other than to give you, as a Azoo user, something else/additional to do while waiting for a set to pass by with a detectable(4 frame's) asteroid. (or to talk about at Azoo, although that's unrelated to tagging)
When it comes to star-bleeds, those seem to be 3rd on the list of most common items in the CSS images (more common, less usefull to tag). And there probably also relative easy to detect code-wise (more easy to code, less useful to tag). On top of that is classifying them in the current classifying page either ambiguous (by only tagging them on one frame) or to much of a click-chore (when tagging them on all frames).
Personally I generally do tag star-bleeds (on all 4 frames) ... But ! that's: Because 1) I use the artifact tag more as a personal tag-note feature to keep track of my image-scanning process. And 2) Because I only have to hit one key to mark a particular location on all 4 frames.
Posted