search

MarsRaw / mars-raw-utils

Utilities for working with publicly available raw MSL & Mars2020 images
MIT License
45 stars 10 forks source link

Curiosity calibration targets. #17

Open plasticalligator opened 1 year ago

plasticalligator commented 1 year ago

image

I am trying to calibrate some colors for the left MASTCAM. Most of the images, akin to above, are covered in dust. Do you know where I could look for images that were taken prior to launch? There are plenty of colored images from NASA but I cannot find any pre-launch raw bayers of the calibration targets.

kmgill commented 1 year ago

The calibration report for Mastcam can be found here: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2016EA000219

You can also check out the calibration data archive on PDS: https://pds-imaging.jpl.nasa.gov/data/msl/MSLMST_0001/CALIB/

plasticalligator commented 1 year ago

Thanks. This was helpful. I managed to figure out what that target (for the ChemCam) is supposed to look like on earth.

image Now I know that the exterior shell is definitely an off-white.

If I can ever manage to get this to a point of completion you (and maybe fellow space monkeys) might be interested in the debayering algorithm I've been working on - but I am a total amateur doing this just as an experiment while teaching myself how camera sensors work on the hardware level so my outputs might be junk.

The following image was not "white balanced" as far as I know and I am unsure why my algorithm outputs the "correct" colors for "white" hence my quest to find images with 'proper' (as if such a thing exists) colors.

image

Github's edit functionality is going to drive me completely insane. The year is 2022 and we don't have functional text editors. The oceans are filled with plastic and we're still using lead pipes for drinking water. Oh well.

Note that red and "blue" are essentially the same in contrast & hue when debayered with your application / malvar.
image That has me believing something other than balancing is going on otherwise the reds should have much more saturation and the "blues" would be closer to blackish.

This is a color-corrected (?) version I found from the NASA website

image

I am also looking at extracting depth from single-frame (without focal stack) images to some degree of subjective outcome. I saw that you added focal stack depth maps as a feature and when I have some energy I plan on looking into how that works.

image

Do you know if Curiosity has any calibration target for stereo images?