View Full Version : Scientifically interesting craters

2012-Feb-14, 03:47 AM
Please tag and discuss images you notice with scientifically interesting craters.

2012-May-07, 08:28 PM
Hi I just analyzed this image with the man versus machine tool and noticed the algorithm did very poorly. After marking for a few minutes I realized that the sun icon on screen was in the lower left corner on the bottom edge, but crater shadows were appearing in the upper left of the craters. If the sun icon is generated from the solar angle used by the algorithm that could explain the poor performance on this image in particular.



2012-May-08, 06:44 AM
Hi John - the sun icon is a completely different thing from the crater detection code.

I run the detection code, which uses template craters that I find within the image and then searches through the image for something that comes close to that. It does NOT do well when the sun is higher overhead (such as the image you show in your post).

Meanwhile, the sun icon is created server-side from a data table that I give to the programmers that uses camera pointing data from the spacecraft. For that particular image, the sun originally came at a 48.32 angle from due East, meaning the sun icon would be in the bottom right; however, to make these appear like depressions to most people, the image was rotated 180 so that the sun comes from 228.32 (roughly the upper left corner, as it looks like to the eye in the image you linked).

In checking the database server-side, the sun angle for that image is recorded as 3.985, but I'm not sure what units that is (it appears in the database to be between -0.8 and 5.8, so I'm thinking it might be in radians ... converting to radians does give an angle of 228.32. So, perhaps there's something wrong with going from the database to display. I'll let the programming team know about this post.

2012-May-08, 02:43 PM
Wow cool. Thanks for the peek inside the science. I saw maybe two or three other images like this yesterday. Didn't affect my mapping, but I noticed poor algorithm performance when the sun icon was off. So probably the images with incorrect icon placement were also high sun angle images.

Thanks for the response and the opportunity to do some citizen science!


2012-May-08, 10:25 PM
P.S. Did the craters look like this thread's image (http://cosmoquest.org/forum/showthread.php?561-Reporting-bad-light-angle)? If so, that could actually be an issue with the craters in the database ...

2012-May-10, 03:12 PM

Spot on. I have seen a few images like that now. When the sun icon is off the craters are way off, like in the link.


2012-May-16, 11:11 PM

Hey team,

I just mapped the crater linked in the man versus machine mode and saw a cool rockslide down the crater wall that I thought you might be interested in. I also see a strange pattern of small craters/boulders along the wall of the crater towards the top of the image as seen on the web page.


2012-May-17, 05:02 PM
Hi JohnMcMillen,

Yes, you are definitely inside a larger crater. The break between the very light area on the left and the slightly darker area to the right is probably where the floor meets the wall.

Neat rockslide there on the upper centre. But, I am not sure which crater/boulder pattern you are referring to in this image.

2012-Jun-03, 10:19 PM
What is this thing? 75

It looks like a geyser inside of a crater, but I don't thing the moon has any geysers so it must be something else.
If I use a little imagination, I can see a spacecraft sitting in the shadows. jk

2012-Jun-04, 04:50 PM
Hi CTroop,

I am not sure what you are referring to exactly in respect to a geyser. But, what you are looking at here is the inside of a mostly shadowed crater.
I hope that helps.

Keep looking....