PDA

View Full Version : Man vs Machine



pstew
2012-Feb-02, 07:55 AM
How far off does a marking have to be from a crater before it is considered to be completely wrong.
Also, there seems to be quite a few images where the makings dont match the craters at all. Should
I delete them all and remark, or try and move the marking to craters.

12

astrostu
2012-Feb-02, 02:59 PM
Hi pstew, it looks like there may have been a mismatch in the auto craters and image in that particular case. If you encounter ones like that, it would be of great help if you could click the "Image Link" directly under the "Done Working" button and post the link here. That way the coders and I can check it in the database and figure out what I messed up.

Once you do that, click your browser's Reload button and get a new image that matches better. Sorry 'bout this! (Does the, "we're still in beta" work for an excuse?)

pstew
2012-Feb-04, 03:10 AM
A list of mismatched crater images

https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00028.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00031.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00107.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00078.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00063.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00152.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00121.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00060.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00098.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00163.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00072.png
https://s3.amazonaws.com/cosmoquest/data/mappers/moon/v2/M111578606RE/1/00178.png

Is that enough or would you like more :)

astrostu
2012-Feb-12, 02:22 AM
Okay, it looks like all of these are from the same original NAC image. This was one that we rotated 180 so that the sun came from the "correct" direction to make it look like you were finding craters, not hills. It appears as though when I ran the automated code, I did not rotate the results by 180, as well. I've gotten a revised list of craters to the programmers and hopefully we'll have those uploaded ASAP.

Thanks for finding this issue!