The semester is starting to come to a close which means, in some sense, that I am rapidly approaching the 25% completion mark of my thesis. Is it too early to already lament the time gone by?
Anywho, the update: research is chugging along or marking time, depending on how you look at it. There have been a number of setbacks. Namely, calibration and shifting of images.
After meeting with Catherine last week, the issue of dark fringes on the CTX images I was using seemed to stick out as odd to her. After digging around, I found out that I had not calibrated the images! Similar to LROC data, the images needed to be calibrated. I initially thought this was the case, but due to completely different functions in ISIS the calibration for CTX slipped under the radar. To illustrate, the commands are
Without the ctxevenodd run on the images, they turned out looking like surface pictures with three large translucent stripes going along them! On top of that, I feel frustrated that these steps aren’t all that clear (but maybe it’s just me). If I, a grad student in planetary science with a background in photogrammetry, can’t make heads or tails of how to preprocess images, how can we expect the public to do so as well? It’s great the information is freely available to anyone, but there is a veil of obfuscation within our community deterring those who may be interested. So all 50 GB of my CTX images need to be redone. It turns out (if you know the right people and have a treasure map to it) there is a site which can do these steps for you. Welcome to PILOT. This, amazingly, will process your images for you and tell you the steps executed. Thank you, PILOT.
Second setback is shifting the images. This is a problem I have been trying to skirt because it’s time-intensive. Due to instrument and photogrammetric constraints, the images will never be exactly in the right spot with 100% confidence. Meaning, the spatial information recorded will have some variance to it, leading to images that aren’t exactly lined up, especially with each other. So, how to fix this? We need large scale and small scale corrections.
Here are a few of the options that have (tentatively) worked for me on a large scale: 1) For HiRISE images, track down some obscure executable script from Livio or visit the annals of the USGS ISIS forums and download it. This will shift the projection system from the associated HiRISE LBL file. 2) For CTX, download the TIFF images from PILOT, which should have the correct(ish) projection 3) For MOLA, I am not sure. I’ve been told not to mess with it in no uncertain terms.
As for the nuanced shifting, in ArcMap you can use:
Copies and permanently moves the raster.
Can take up to an hour to do!
Adjusts the auxiliary files spatial information.
Can take a minute or two, but isn’t always functioning because of mismatched Datums.
Georeference Tie Points
Warps one image to another.
Can be good for angled images.
Introduces another subjective error and can take around 20 mins to do.
Further, you will need to locate the nadir-most HiRISE image and shift everything (except MOLA) according to that one. This can get quite tedious with a project like Mojave, with ~15 HiRISE images and 7+ CTX images!
Another part of my daily routine lately has been data management. My local machine has 232GB of total space, which is shared with my peers. Projects can range from 1GB to 40GB. Learning to prioritize which data to hold onto and which to place on the lab’s 4TB server for storage has been something new. I have liked the challenge, but it can be tedious sometimes.
Even with all of this complaining, this semester, hands down has been one of the best in my college career. I am so excited for the next steps, each with their own quirks and challenges. Walking up to each stair and questioning its existence and purpose. Iconoclastic or foolhardy? Stay tuned to find out. All in all, I can really see myself as part of this community going forward and I am so thankful for the opportunities I have had to date.