IPB

Welcome Guest ( Log In | Register )

37 Pages V  « < 9 10 11 12 13 > »   
Reply to this topicStart new topic
KBO encounters
centsworth_II
post Jul 12 2011, 04:59 AM
Post #151


Senior Member
****

Group: Members
Posts: 2150
Joined: 28-December 04
From: Florida, USA
Member No.: 132



Even some very faint objects aren't getting past Icehunters' scrutiny.
Attached Image



Some of the less subtle objects also capture attention -- for less than scientific reasons.

When I saw this:
Attached Image


I thought of this:
Attached Image
Go to the top of the page
 
+Quote Post
stewjack
post Jul 12 2011, 06:17 PM
Post #152


Member
***

Group: Members
Posts: 252
Joined: 5-May 05
From: Mississippi (USA)
Member No.: 379



QUOTE (centsworth_II @ Jul 11 2011, 11:59 PM) *
Even some very faint objects aren't getting past Icehunters' scrutiny.



I mark objects like that all the time, but I rarely get more than one other person to agree with me. In fact that may be one of mine! You could argue that they are "grey" not "white," but if I click on the "change contrast" button, and the object looks white - then I mark it. However, it must remain "blob like," and not become square-ish or pix-elated with the increased contrast.

163 found in
324 images

Jack
Go to the top of the page
 
+Quote Post
Greg Hullender
post Jul 17 2011, 04:01 AM
Post #153


Senior Member
****

Group: Members
Posts: 1017
Joined: 29-November 05
From: Seattle, WA, USA
Member No.: 590



I just finished my 1000th screen, so I thought I'd make some more-specific suggestions to make the tool easier to use.


1) As I said before, I want to press "space" to advance to the next screen instead of having to click a button with the mouse. If I press space without having marked any objects, that means I think there's none there.


2) Also like before, I'd like to press "b" to advance to the next screen AND signal that something was wrong with the current one. The new twist is that I may have clicked some objects anyway. That makes sense; many of the images aren't entirely wrecked, but "b" signals that I think they need to retake the picture.

3) Instead of an asteroid button, I'd like it if I could just click once to circle something, then click inside the circle to change to an asteroid, and then click a third time to erase.

4) I wish the application would start loading the next screen in the background so I didn't have to wait so long.

5) I wish I could go back one screen; sometimes I see an object AFTER I've clicked next, but there's nothing I can do about it.

Does anyone know how to contact the team that manages the app? I looked all over the site (and posted some suggestions there) but I didn't find an obvious way to send feedback.

--Greg


Go to the top of the page
 
+Quote Post
ElkGroveDan
post Jul 17 2011, 04:12 AM
Post #154


Senior Member
****

Group: Admin
Posts: 4750
Joined: 15-March 05
From: Sloughhouse, CA
Member No.: 197



I agree with #5.


--------------------
If Occam had heard my theory, things would be very different now.
Go to the top of the page
 
+Quote Post
dilo
post Jul 17 2011, 08:05 AM
Post #155


Senior Member
****

Group: Members
Posts: 2492
Joined: 15-January 05
From: center Italy
Member No.: 150



I agree too, especially on last 3 points!


--------------------
I always think before posting! - Marco -
Go to the top of the page
 
+Quote Post
nprev
post Jul 17 2011, 08:13 AM
Post #156


Senior Member
****

Group: Admin
Posts: 8406
Joined: 8-December 05
From: Los Angeles
Member No.: 602



I aqree with #5 as well; had that happen several times, and it just kills me!!!


--------------------
A few will take this knowledge and use this power of a dream realized as a force for change, an impetus for further discovery to make less ancient dreams real.
Go to the top of the page
 
+Quote Post
NickF
post Jul 18 2011, 03:32 AM
Post #157


Member
***

Group: Members
Posts: 107
Joined: 29-January 09
Member No.: 4589



QUOTE (Greg Hullender @ Jul 16 2011, 11:01 PM) *
5) I wish I could go back one screen; sometimes I see an object AFTER I've clicked next, but there's nothing I can do about it.


I flick the mouse pointer away from the 'done marking' button after I've clicked it. That stops my 'click throughs'.


--------------------
Protein structures and Mars fun - http://www.flickr.com/photos/nick960/
Go to the top of the page
 
+Quote Post
centsworth_II
post Jul 19 2011, 10:27 AM
Post #158


Senior Member
****

Group: Members
Posts: 2150
Joined: 28-December 04
From: Florida, USA
Member No.: 132



QUOTE (Greg Hullender @ Jul 16 2011, 11:01 PM) *
5) I wish I could go back one screen...
I feel y'all's pain, believe me. I've missed more objects than I care to admit. But that's part of the "game".
From IceHunters FAQ:
Attached Image


You may agree with their reasoning or not, but there it is. It is possible to come to terms with -- even be at peace with -- the lack of a "back" button, but it involves a denial of the human ego, setting aside any perfectionist tendencies, and embracing the notion that you are but one of an army of worker ants moving toward a communal goal.
Go to the top of the page
 
+Quote Post
stewjack
post Jul 19 2011, 04:08 PM
Post #159


Member
***

Group: Members
Posts: 252
Joined: 5-May 05
From: Mississippi (USA)
Member No.: 379



NOTE: This post is only for IceHunters or people interested in IceHunters. Perhaps we should start a new thread?

QUOTE (centsworth_II @ Jul 19 2011, 05:27 AM) *
-- an army of worker ants moving toward a communal goal.

IMO: That is the secret of this science for dummies, (ie. human-beings). project. Accuracy takes time and effort. Even highly trained or dedicated people can get bored or distracted. With IceHunters, every image is checked by a minimum of eight people, and they overlap images, so I suspect each area of your image is checked by a minimum of 16 people! Normally, within about five minutes of clicking the [Done Marking] button eight or more people have completed the image and the results are reported.

Note: Sometimes overlap image results are also reported at the same time, but I don't fully understand what I am seeing. This is because the data from overlap images is not labeled clearly. It is only obvious that overlap image results are being reported when the same object is reported twice by two separate groups of three or more individuals. It seems to me, that would have to be from two different images.

Personal Example: When I compared my identified objects against a large number of final group results, I found that I was missing about one object for every 9 or 10 objects the group found. Also I was reporting about 2 objects that those same group results had never noticed or at least accepted as possible KBO's.

I find these result acceptable.especially for missed objects. I would be interested other peoples reasoning.

Note: At the present time, only ~28% of the objects that have been reviewed are found acceptable for follow-up by the reviewer.

Jack
Go to the top of the page
 
+Quote Post
Greg Hullender
post Jul 20 2011, 03:35 AM
Post #160


Senior Member
****

Group: Members
Posts: 1017
Joined: 29-November 05
From: Seattle, WA, USA
Member No.: 590



QUOTE (centsworth_II @ Jul 19 2011, 02:27 AM) *
You may agree with their reasoning or not, but there it is.

I didn't realize they had done it on purpose. I've supervised data collection and labeling efforts at four different companies over the past thirty years, and, in my experience, lack of a back button is a mistake. The reason is, people are uncomfortable with an irreversible decision. They go *faster* if they have the option to go back--even when they never actually use the button.

That said, this is a much, much faster task than most of the kinds I've supervised, so the user has a lot less to lose from a mistake. Accordingly, it might not make as much difference here. I don't believe it would make things worse, but the positive benefit might be small in this particular case.

Far and away the biggest problem is the lack of keyboard shortcuts, which makes it an RSI hazard. Everything else is really secondary.

--Greg
Go to the top of the page
 
+Quote Post
centsworth_II
post Jul 25 2011, 04:16 PM
Post #161


Senior Member
****

Group: Members
Posts: 2150
Joined: 28-December 04
From: Florida, USA
Member No.: 132



In response to a request by the science team, Icehunters are really getting serious about finding faint objects!
Attached Image
Go to the top of the page
 
+Quote Post
Greg Hullender
post Jul 26 2011, 12:04 AM
Post #162


Senior Member
****

Group: Members
Posts: 1017
Joined: 29-November 05
From: Seattle, WA, USA
Member No.: 590



Yes, and now it makes sense to use the contrast button. It really makes the faint objects pop!

--Greg
Go to the top of the page
 
+Quote Post
centsworth_II
post Jul 27 2011, 01:50 PM
Post #163


Senior Member
****

Group: Members
Posts: 2150
Joined: 28-December 04
From: Florida, USA
Member No.: 132



By the way, for anyone here that may be Icehunting but not reading the Icehunter forums, here are examples from the science team. I still mark the bright ones because its good to get all the variable stars into the data base (and who knows, it really could be a huge KBO!).
Attached Image
Go to the top of the page
 
+Quote Post
tfisher
post Jul 30 2011, 02:30 PM
Post #164


Member
***

Group: Members
Posts: 204
Joined: 29-June 05
Member No.: 421



Does anyone know how far a genuine KBO is likely to have moved in the icehunter images? I've been trying to watch for white/black blob pairs. After 1000+ images, my current favorite bet in the objects I've marked to be an actual KBO is this black-white pair (as seen in positive and negative -- I got both smile.gif )

IH-X-515823
IH-X-515965

I'm curious, though, for the image pairs being subtracted what is the range of motion for an actual KBO between the images?





Go to the top of the page
 
+Quote Post
john_s
post Jul 30 2011, 03:21 PM
Post #165


Member
***

Group: Members
Posts: 580
Joined: 3-December 04
From: Boulder, Colorado, USA
Member No.: 117



Pairs can be anywhere from a few hours to a few days apart, and at opposition typical KBO motions are 3 arcsec/hour or 15 pixels/hour. So motion is likely to be tens to hundreds of pixels.

Sometimes an apparent pair results from a CCD defect that is fixed on the camera, but moves slightly when the images are shifted to line up the stars. It's possible that's what you're seeing here. Or, it *might* be a KBO if this pair was taken unusually close in time.

By the way, we just got the first 2011 data through the differencing pipeline, and expect the new images to be posted on the Zoo very soon...

John

Go to the top of the page
 
+Quote Post

37 Pages V  « < 9 10 11 12 13 > » 
Reply to this topicStart new topic

 



RSS Lo-Fi Version Time is now: 12th December 2019 - 02:28 AM
RULES AND GUIDELINES
Please read the Forum Rules and Guidelines before posting.

IMAGE COPYRIGHT
Images posted on UnmannedSpaceflight.com may be copyrighted. Do not reproduce without permission. Read here for further information on space images and copyright.

OPINIONS AND MODERATION
Opinions expressed on UnmannedSpaceflight.com are those of the individual posters and do not necessarily reflect the opinions of UnmannedSpaceflight.com or The Planetary Society. The all-volunteer UnmannedSpaceflight.com moderation team is wholly independent of The Planetary Society. The Planetary Society has no influence over decisions made by the UnmannedSpaceflight.com moderators.
SUPPORT THE FORUM
Unmannedspaceflight.com is a project of the Planetary Society and is funded by donations from visitors and members. Help keep this forum up and running by contributing here.