PDA

View Full Version : Cheating or good technique?



imported_Thumper
2012-May-02, 02:42 PM
Using latest Firefox browser with MS VISTA, NVIDIA color .
After some difficulty I finally settled on the best color/hue/contrast setting for balancing the Viewsonic CRT monitor button controls and the NVIDIA color Panel adjustments for general crater marking (for my eyes and room lighting so would be no help to others to post that).
But for the 'cheating' part: I use a screen zoom level that fills a 23" screen with just the normal gray-tone moon image and marking controls. After marking all the potential craters found that way, I discovered that if I placed the cursor just past the edge of the image, held a left click and dragged the cursor back onto the image, the (somewhat) dull gray-tone changed to a much brighter blue-tone image. While this is just the 'highlighting' of the SELECT Mode, the resulting change of contrast sometimes allowed picking out additional faint features. It aided for finding the older 'weathered edges' crater locations. Then tap the mouse key to return to normal mode to mark the crater. Has anyone else used this or other easy image enhancement techniques?
So the 'cheating' question: Is it better for the project to maybe add some additional older faint craters to the totals or does this technique mess with your determination of finding the best illumination angles and image contrast for future image identification programs?

Would it be possible/useful to add software program controls for the user to change image overall color, or maybe a flip to a negative image? Just joined on 1 May. Thanks.

kmasterdo
2012-May-02, 06:33 PM
Welcome to CosmoQuest, Thumper! I'll point the science team to your question.

astrostu
2012-May-02, 06:40 PM
Hi Thumper - your method is not cheating, though I think very few people will do it. When experts identify craters, we all use a wide array of techniques, but almost all of them allow us to actively zoom in and out and adjust contrast and brightness. Our interface on MoonMappers doesn't allow this because we wanted to keep it as simple as possible, but you doing this yourself is perfectly fine.

If I'm using the interface, I'll often CTRL+(mouse scroll up) (I'm on a Mac) which zooms the screen in to over-sample the image. A bigger image on-screen makes it easier for me to be more precise.

The only words of caution would be to make sure when you do this, you aren't just bringing out artifacts. I almost entitled my doctoral dissertation (where I identified 700,000 craters on Mars, by hand): "Martian Crater Identification: An Exercise in Pareidolia (http://en.wikipedia.org/wiki/Pareidolia)."

imported_Thumper
2012-May-03, 12:22 AM
Thanks. It probably wouldn't matter much anyway as my learning curve looks like a doodlebug trail - bouncing between over and under identifying craters.