Slashdot Mirror
Curiosity's Latest High-Res Photo Looks Like Earth
New submitter bbianca127 writes "Curiosity sent a picture down to us, and it looks a lot like Earth. Actually, the picture's color quality has been changed — to human eyes, the landscape would look a lot more reddish. Still, it looks remarkably like the southwestern United States (bringing to mind the Arrested Development quote about how Lucille Bluth would rather be dead in California than alive in Arizona)." Definitely a different sense of the place than the one given by the reddish-brown posters I remember from elementary school.
Here is a page on the MSL's site where you can see both versions of the photo:
http://mars.jpl.nasa.gov/msl/multimedia/images/?ImageID=4431
One is white-balanced and one colored. The white-balanced version represents what the scene would look like to human eyes under an Earth sky. The colored represents what the scene would look like to human eyes on Mars.
The point of using white-balanced photos is that geologists are used to looking at rocks on Earth. So when a geologist wants to judge rock characteristics using color, it helps to white-balance it so the color is similar to what it would be if looking at those rocks on Earth.
__
Big apple, new Yorik, undig it, something's unrotting in Edenmark.
Science.
Because NASA doesn't have advertising?
(Which is a completely wasted opportunity.)
Yyyyeah...they're not "altering" the photo. What they're doing is balancing the color so that people can know what they are seeing. The reason for this is that the Martian atmosphere has radically different color properties from that of our own. What this means is that visible observations cannot be made reliably: for example, a red rock on mars may not actually be red as we understand the color, and so conclusions geoloists make based on a color may be erroneous, because they are basing those conclusions on colors observed under earth's sky.
If anyone's interested, another scene is shown with and without white balance here: http://apod.nasa.gov/apod/ap120815.html
As much as I love the awesome idea of moving a chunk of terrain between planets, I'm going to shoot for an informative mod and answer the question.
There is a sundial mounted on Curiosity, with a few colored stripes on it. Those stripes' colors (red, green, blue, and yellow) were recorded under Earth's lighting, Now that those same stripes are on Mars, their apparent color change in new pictures is the result of Mars' different lighting. By comparing the stripes' pictures, an approprite transformation can be determined, then applied to other pictures to compensate for the change in lighting.
We are sure because we're assuming that those stripes' actual colors haven't changed significantly during flight or landing.
You do not have a moral or legal right to do absolutely anything you want.
Not really...
If you don't mind being unable to take color shots of relatively fast moving things, you can use a conventional greyscale sensor, swap color filters between frames, and then crunch the result into a color image(or, if you have the space and don't mind a moderately complex optics package, you can have three greyscale sensors, each with a fixed color filter). If you need a color image within one frame, you use a fixed bayer(or similar) filter and demosaicing. Eats nontrivial resolution compared to the pure greyscale or swapped filters strategy; but you get everything in one shot and fewer moving parts. Then you have the somewhat oddball Foveon approach, where your greyscale sensors are stacked vertically, and use the different rates of absorption in silicon of different frequencies to do the filtering...
In very broad terms, they all have the 'greyscale sensors and filters' strategy; but there are a fair few ways to go about it. If you count chemical and biological sensors, you are more likely to find sensor elements that are actually tuned to a specific wavelength, rather than filtered to it; but the final image is still a matter of crunching together results from individual elements that are really only giving you intensity data for a relatively narrow slice of frequencies.
On the rover are color calibration targets (here is the one for the rover's arm's instruments). We know exactly what the colors of those targets are supposed to look like, when imaged by the cameras on the rover, under normal Earth-like lighting conditions. By looking at how those targets appear in the images we get back under Mars lighting conditions, we can do two things:
1) Learn a lot about the lighting conditions on Mars.
2) Correct the appearance of images we get back to correct for that Mars lighting.
Obviously this must mean that Martian rocks and Earth rocks share a common ancestor!
Yes. It does. That common ancestor is called the protoplanetary disc which led to the formation of the inner solar system.
Now go troll somewhere else.
Entia non sunt multiplicanda praeter necessitatem.
Gee, I wonder if such an image could be available on NASA's web site. Nah, that's unthinkable.
Oh, wait, here it is: http://mars.jpl.nasa.gov/msl/multimedia/images/?ImageID=4431
No sig today...
Here you go:
Original:
http://photojournal.jpl.nasa.gov/jpeg/PIA16051.jpg
White balanced:
http://www.nasa.gov/images/content/676031main_pia16051-figure_2_brightened-portalfull.jpg
You would, but geologists wouldn't. They are used to what rocks and minerals look like under our own earthly lighting. As such it makes sense to adjust the color of the image to match earth-normal lighting conditions.
Begs the question is more properly used when I say "So now that you stopped beating your wife, how is your marriage?" That begs the question of if you ever beat your wife at all.
The poster asking about a robot and a mirror should have used "Raises the question" instead.
Even on earth we have this issue (I've made a fairly healthy living navigating through color space to color space and light source to light source over the years). People seem to forget that our own sunlight can vary during the day, geographical location, cloudy days, etc, and indoor lighting is the beast with a billion backs. Even your own eyes can betray you, needing a moment to adjust, and often one eye sees color slightly differently from the other.
Color scientists have had an absolute color and light source standard to measure against (CIE LAB) or 40+ years; Mars (or anywhere in the universe that receives light in the visible spectrum) fits just dandy into this model for color transformations, it's just a bit further away than usual. The less light there is to measure, the smaller the total color gamut will be, but you can extrapolate pretty well, if you don't mind some +/- errors along the way.
Typically, a true simulation would need several hundred color swatches for analysis, plus an iterative scanning approach to nail down the color gamut points that are furthest away (say, blues could be further off than reds, so require more attention for a transform). Still, for a general "this is approximately how it'd look on Earth" a 4 swatch RGBY spectrum is close enough.
It's something like the difference of having a precision of tenths to a precision of hundred-thousandths, when all you're doing is counting apples. You may be plus or minus a tenth of an apple, but so what?
The only thing that's a little surprising is that they didn't include a calibrated black strip, but I suppose they didn't really need to account for the variation between deep shadow areas or very dark objects in this case.
Actually, you probably would see it more like the white balanced photo than the regular one. Your brain is very good at auto white balance.
Perceptual re-balancing is very different from absolute colorimetric re-balancing, which is what is used here.
A late evening shot (which this basically is) looks very different when you balance it against a Gretag Macbeth card than if you balance it according to human perception.
NASAs goal here is clearly to make the picture as useful as possible to those who study them, not to give the public a "true" image of what we would perceive if we were there. I think there should be room for both.
When the spectrum of ambient light does not match that of "white" light (which is simply the particular spectrum we evolved to perceive), the eye's photoreceptors become disproportionately fatigued, and perception of the light's color drifts toward white. You can experience this phenomenon yourself if you light a room entirely with red party lights. Soon, your red photoreceptors will become fatigued and the colors of objects in the room begin to appear more normal. I think explorers on Mars would experience the same effect. So photos like this are actually how it would look to them.
The atmosphere is so thin it's basically vacuum, so the view of the stars should be pretty good. If we could engineer cottonwood trees that thrive in vacuum, high radiation, temperatures as low as -150 celcius, and no water, we'd be good there too. Of course then we'd have to engineer humans that didn't suffer bone decalcification due to the low gravity...
Snarky as your comment may have been meant, I think you need to check your numbers again what constitutes "so thin it's basically vacuum."
Mars has an average surface atmospheric pressure of 0.636 kPa. Earth has 101.325 kPa. So yes, while it is 160-times thinner, that's still pretty thick, especially if dust is kicked up. After all, remember that with 1/3rd gravity, much less air friction and no moisture, dust particles can stay afloat for quite some time.
And then, compare that to the moon, with a pressure of 10^-7 kPa (~1 nPa), Mars still has a 6.36 million times denser atmosphere. And compared to interplanetary space, that's still practically solid, as space has 400.000 times less pressure.
In other words: If Mars is a near-vacuum at nearly 10^17 times more molecules per cm than interplanetary space, then a snail that moves at only 3*10^10 cm/s.