Nature of project: experimental, data analysis
Available to students on full-time physics degree schemes or joint students.
The aim here is to use similar illumination imagery of the Moon at 1 metre resolution, taken on different dates, to look for changes caused by fresh meteorite craters, landslides, or as the result of rare but violent magnitude 3-5 shallow depth moonquakes.
You will do this utilizing NASA Lunar Reconnaissance Orbiter Camera (LROC) images taken many months/years apart. These will need to be morphed to one another for registration purposes, and then normalized in brightness to one another using United States Geological Survey ISIS stereo matching software, or a modified NASA JPL stereo matcher. Correlation based stereo matching between the temporal image pairs will reveal any significant differences which could result from one of the above forms of geological activity. Please expect to find quite a few false candidate detections of change, for example cosmic rays radiation events on the CCD - so you will have to do some manual checking of the results to decide which are lunar surface changes and which are not.
As correlation based stereo matchers work typically over a fixed correlation patch size e.g. 10x10 pixels, it is probably worth working with reduced resolution versions of the images so that you can investigate changes on different spatial scales.
Once you have found some candidate changes, you should check for other images covering the area to see if you can narrow down when the change happened. If shadows are present you can find the slope angle and see if changes are more common on slopes, as one would expect for landslides.
A successful project will develop beyond the above in one/some of the following directions:
1) Learn how to use United State Geological Survey ISIS software to process NASA LRIC images
2) Identify overlapping image pairs with similar illumination, but taken on different dates
3) Use stereo matching software to find corresponding pixels in the 2nd image for every pixel in the first image
3) Warp the 2nd image to the first, normalize one to the other and subtract
4) Look for and quantify differences found either as a result of a bad stereo matcher correlation between images, or by manually blinking between images
When considering where to take your project, please bear in mind the time available. It is preferable to do fewer things well than to try many and not get conclusive results on any of them. However, sometimes it is useful to have a couple of strands of investigation in parallel to work on in case delays occur.
Additional scope or challenge if taken as a Year-4 project: It would also be interesting to demonstrate the detection of sunglint from debris left behind by successful and crashed landings on the Moon. An example has been found already of Sunglint at the Apollo 14 landing site, from some of the gold foil heat insulation blanket that was strewn around the landing site during takeoff.
A more interesting location to study would be around the predicted impact site of ESA's SMART-1 mission. This has never been found, and ESA are offering a bottle of Champagne to anyone who can conclusively find this! The sunglint method might be the only way to do this as we do not have any "before" impact images to compare modern images against.
Please speak to Tony Cook if you consider doing this project.
Initial literature for students:
It would help if you had some experience/enthusiasm for image processing.
Assistance will be provided by your supervisor over image formats, cartography etc.
|milestone||to be completed by|
|Research out stereo matches and decide which one to use||end of October|
|Become familiar with spacecraft lunar imagery and select some candidate temporal image pairs||Christmas|
|Extensive stereo matching runs, and collation of candidate change locations||end of February|
|Rating and analysis of candidate change sites||Easter|