3

A paper from the 2017 meeting Lunar and Planetary Sciences XLVIII describes a systematic observation of the Martian sky, with a sequence of eight taken every day at zenith and near-horizon angles. They show tenuous clouds, often with a wavy gravity wave pattern (that's gravity wave, not gravitational wave).

The images are very grainy compared to the one's I'm used to seeing of the Martian sky or surface. Are they taken before sunrise when it is very dark? Since nothing is moving couldn't they just use a longer exposure? Or does it have to do with noise from some enhancement processing? If so, what is the enhancement process?

Isn't extensive cloud cover quite rare on Mars, and not a daily occurrence?

Note: the GIF below is made from only four frames of the first 8-frame GIF of three, shown in this news item in Science. They look much better there - the stackexchange image size restriction prevents the direct posting of the 2.7 MB images. You can direct-link to them: here, here, and here.

enter image description here

For contrast, here is a tweeted image.

enter image description here

uhoh
  • 148,791
  • 53
  • 476
  • 1,473
  • 2
    I guess the structures are so faint, you hardly can see them at all. You have to increase contrast quite a lot to get images like that. The 'grain' mostly looks like compression artifacts which are pronounced by this as well. Longer exposure doesn't help as there is already enough light on the sensor, but differences between pixels are just too low. – asdfex Mar 23 '17 at 11:27
  • @asdfex A longer exposure would certainly help if you read out the CCD multiple times during the exposure, à la the way astrophotography is done. Maybe these are already so long that it would cut into Curiosity's work schedule? I'm curious what their exposure time actually is. Is ti before sunrise? – uhoh Mar 23 '17 at 11:31
  • @asdfex If it is in fact shot noise, the improvement would only scale as the square root of total time, so maybe it is just a practical operational limit. – uhoh Mar 23 '17 at 11:44

1 Answers1

5

The Martian atmosphere is rather thin, and opacity is rather low. The sensitivity (ISO) needs to be cranked up to see the Martian sky at all. This alone will create some amount of graininess. Even with that, from the linked Science article, "The captured clouds are so thin as to be invisible without painstaking computer enhancement, Kloos said in his presentation Tuesday."

This almost certainly means a good amount of contrast enhancement such as histogram equalization, and perhaps sharpening. Both will enhance contrast at the expense of enhancing noise.

David Hammen
  • 74,662
  • 5
  • 185
  • 283
  • It looks like shot noise to me. Why not use a longer total exposure/integration time like any other "space camera" would? Is the total integration time already very long? – uhoh Mar 23 '17 at 14:40
  • 2
    I cannot find information on the time between each of these frames, but extending the exposure time could introduce motion blur depending on how quickly the clouds are changing. – OrangePeel52 Mar 23 '17 at 16:24
  • @OrangePeel52 That's a very good point! I hadn't thought of that, but it would be a hard limit. One can de-motion-blur ground images from a narrow FOV camera or telescope in orbit by reading out frequently and shifting/correlating the images before summing, but not this kind of image. Yep. I don't know if this the reason or not, but it certianly is one limit to the total integration time. – uhoh Mar 23 '17 at 22:50
  • @David, opacity is about 0.5, so not too bad. Note that you can easily see the sky in all the panorama images. It's just a problem of contrast, nothing to do with ISO. – asdfex Mar 24 '17 at 12:31
  • @asdfex I still haven't figured out the time, and solar elevation during these exposures. I think the water vapor is visible when the temperature is low - are these taken at twilight - before the sun actually rises? Remember those panorama images you mention are taken during the day specifically because that's when there is a lot of light. – uhoh Mar 26 '17 at 07:27
  • @uhoh the article you linked says "early morning clouds' in the caption of the image you show above. But the image is clearly taken after sunrise, because you can see the terrain in the lower left corner and shadows of the clouds are visible. Morning and evening are preferred times for such images because you get better contrast due to shadows within the clouds. – asdfex Mar 26 '17 at 13:05
  • @asdfex That's a very subjective interpretation of a grainy, black and white image of a world you've never been to! I'll wait for some credible, official source for the the image specifics. Have shadows on the surface produced by atmospheric clouds ever been reported on Mars? Can you cite a source? – uhoh Mar 27 '17 at 08:15
  • Well, if you can see them (with a certain level of contrast), they have to produce shadows on a related scale. The statements I made are the same no matter which planet we talk about. – asdfex Mar 27 '17 at 12:18
  • @asdfex No.not necessarily. For pre-dawn images lighting would be diffuse, and there would not be any distinct or recognizeable shadows. Terrain color, texture, and orientation will produce far more visible contrast than shadows cast by essentially invisible, extremely tennuous clouds. – uhoh Mar 28 '17 at 14:24
  • Not after the kind of image processing applied here - a huge amount of normalization seems to have been done, keeping only local but not global contrast. – asdfex Mar 29 '17 at 11:07
  • David, @asdfex , I've just added an image to the question. I'm not sure what it implies, but it looks a lot nicer than the others, and it's from the Opportunity rover, not curiosity! – uhoh Apr 03 '17 at 07:27