NAVCAM’s shades of grey

Ever since early August, when Rosetta rendezvoused with Comet 67P/Churyumov-Gerasimenko at a distance of roughly 100 km, the on-board navigation camera (NAVCAM) has been returning images that depict the many different facets of its nucleus. A complex surface emerges from these images, revealing valleys, cliffs, boulders, and craters all over the comet.

NAVCAM takes black-and-white images and the surface of 67P/C-G shows a wide dynamic range of light and dark regions, depending on the illumination conditions and surface characteristics at any given area. But what do “light” and “dark” mean for an object like 67P/C-G? Followers of this blog have asked this and similar questions, so here are some details on how NAVCAM images are taken and displayed to make a wide range of surface features possible.


Four-image NAVCAM mosaic of Comet 67P/Churyumov-Gerasimenko, using images taken on 24 September 2014 when Rosetta was 28.5 km from the comet. On the left, the contrast was enhanced by setting the darkest pixels as black and the brightest ones as white; on the right, the intensities were scaled so that the mean brightness of fully illuminated regions of the comet is around 4%. Credit: ESA/Rosetta/NAVCAM

Let’s start with the light available to take pictures by. At present, 67P/C-G and Rosetta are out beyond the orbit of Mars and the Sun is roughly only 10% as bright as they would see if they were in orbit around the Earth. In addition, the surface of comets can be very dark, reflecting less than 10% of the light that falls on them – something that has been known since ESA’s Giotto flyby of Comet 1P/Halley in 1986. The technical term used is that comet nuclei have a very low ‘albedo’. For 67P/C-G in particular, astronomers have combined visible light data from the Hubble Space Telescope and ground-based observatories, with infrared data from Spitzer and WISE, to determine that it has an albedo of just 4–6%, as dark as charcoal.

So combining these two facts, there’s not that much light coming from 67P/C-G with which to take a picture. But just as you would do in dimly lit situations on Earth, that can be overcome by using a longer exposure time. In particular, the exposure time needs to be long enough to get above the background noise of the detector, but not so long that any parts of the scene saturate the detector. With NAVCAM, the aim is to get the brightest parts of the comet up to roughly 75–85% of the detector saturation limit, which at present means an exposure time of 6 seconds.

Once a NAVCAM image has been captured and sent back to Earth, it is processed to remove artefacts due to electronic noise. The data are then scaled for display according to their brightness: if left untouched, the darkest parts in the image, where there is essentially no light, will be black, while the brightest parts will be at about 75–85% grey (where 100% grey is white). In practice, some slight additional tweaking of the brightness and contrast is done to bring out the full range of features, with the result that the brightest parts of the nucleus are just about white.

While this is a perfectly standard approach, it admittedly doesn’t give a completely accurate impression of the physical nature of the comet, where even the white parts of the picture are in fact very dark.

But actually, the human eye and brain do this all the time, constantly adjusting their sensitivity and perception of intensities to the scene at hand or even locally within a given scene. This so-called “anchoring” effect is the reason why the Moon appears white or even to shine against the dark night sky, while we know – not least from photos shot by moonwalking astronauts, as well as direct measurements of samples of the lunar soil – that the Moon’s surface is in fact a dark shade of grey, with an average albedo of around 12%.


The “checkershadow” illusion, an optical illusion published by Edward H. Adelson in 1995; on the left hand side, squares A and B look to have very different brightness, but are in fact the same, as demonstrated on the right. ©1995, Edward H. Adelson, Massachusetts Institute of Technology

A classic example is provided in the “checkershadow” optical illusion, which shows a column casting a shadow over parts of a checkerboard pattern. Because of the anchoring effect, the human eye perceives the squares in the shadow to be lighter than they really are, while those outside the shadow are seen to be lighter. In fact, squares A and B are exactly the same shade of grey, which can easily be shown by masking out the rest of the picture.

So, a human in a spaceship next to 67P/C-G may in fact perceive the comet pretty much as it is seen in the intensity-stretched NAVCAM images, if not even brighter. But one way of giving at least a suggestion of just how dark comets are is to show 67P/C-G against a number of other Solar System objects exhibiting a wide range of albedos.

The montage below compares 67P/C-G with the Moon, the Earth, and Enceladus, a moon of Saturn. The brightness of each object in the montage has been scaled according to its mean albedo: for 67P/C-G, we have taken an albedo of 5%; for the Moon, 12%; for Earth, 31% (with deserts having an albedo of roughly 40%, thin clouds 30–50%, thick clouds 60–90%, and oceans 7–10%). Finally, for simplicity, an albedo of 100% has been taken for the brightest parts of the ice-covered surface of Enceladus, the most reflective body in the Solar System.

It’s hard to do this scientifically accurately, partly because the actual albedo in a given image of an object depends on a whole host of factors and because the human eye and brain don’t respond linearly to different light levels. But hopefully this comparison gives at least an impression of quite how dark 67P/C-G is and how diverse the Solar System’s bodies can be.


Images of Enceladus, the Earth, the Moon, and Comet 67P/C-G, with their relative albedos scaled approximately correctly. The albedos that were used are: 100% for Enceladus, 31% for the Earth, 12% for the Moon, and 5% for 67P/C-G. The images are not however to scale physically. The image of Earth was taken by the OSIRIS camera on Rosetta during the spacecraft’s last flyby past our planet in 2009.
Credit: NASA/JPL/Space Science Institute (Enceladus); ESA/Rosetta/MPS for OSIRIS Team MPS/UPD/LAM/IAA/RSSD/INTA/ UPM/DASP/IDA and Gordan Ugarkovich (Earth); Robert Vanderbei, Princeton University (Moon); ESA/Rosetta/NAVCAM (67P/C-G).

Another interesting aspect of the NAVCAM images of 67P/C-G is the very crisp and deep shadows. This is a result of a single, distant illumination source, namely the Sun, and the lack of any atmosphere surrounding the comet which would serve to diffuse the light, as on Earth: the developing coma is far too thin to scatter much light. However, on occasion, the deep shadows may be relieved by light reflecting off other parts of the comet, as seen in a number of the NAVCAM images.

Beyond the overall darkness of 67P/C-G, why are some regions nevertheless lighter and darker than the average? Some of this will likely be due to compositional differences across the surface of the comet, with some regions fresher due to activity and others more deeply covered in dust. The Rosetta scientists are studying the composition of the surface through a combination of imaging at different ultraviolet, visible, infrared, and millimetre wavelengths, along with a range of other remote-sensing diagnostics.

But one key difference is down to the angle between the incident sunlight hitting the comet and the reflected light being measured by Rosetta, which scientists call the ‘phase angle’. When the phase angle is large, the spacecraft sees shadows cast over the surface by the light coming in at an angle, but when the phase angle is small, few shadows are seen.

Asteroid Steins seen from a distance of 800 km, taken by the OSIRIS imaging system from two different perspectives during the flyby.  Credits: ESA 2008 MPS for OSIRIS Team MPS/UPD/LAM/IAA/RSSD/INTA/UPM/DASP/IDA

A montage of images of asteroid Steins from Rosetta’s flyby in 2008. In the first image (top left corner), the asteroid was was right ahead of Rosetta while the Sun was at the spacecraft’s back. In this configuration, with the phase angle close to zero, the ‘opposition surge’ effect is evident: the asteroid appears very bright and hardly any shadows are cast. In the following images, as the spacecraft moved and the phase angle increased, the overall brightness is lower but shadows are present, revealing more features on the surface of the asteroid. Credit: ESA 2008 MPS for OSIRIS Team MPS/UPD/LAM/IAA/RSSD/INTA/UPM/DASP/IDA

As Rosetta has been manoeuvring around 67P/C-G, we have been seeing different regions at a wide variety of phase angles and because 67P/C-G has such a complex shape, different regions are seen at different phase angles simultaneously, further enhancing the apparent diversity of its landscape.

A particular effect occurs on many airless bodies in the Solar System whose surfaces are covered in dust, including the Moon, asteroids, and comets, when the phase angle reaches zero: with the light coming from behind the observer, a significant brightening is seen, the so-called ‘opposition surge’. This is in part due to the fact that, under these conditions, there are no shadows visible anymore: they are all ‘hidden’ behind the objects (e.g. dust particles and rocks) being illuminated, as seen from the observer. Another effect may occur if the dust particles are roughly the same size as the wavelength of light being measured: this can result in them acting as little retro-reflectors, again causing an increase in brightness at zero phase angle.

Thus, despite the very dark nature of the comet, there is nevertheless much to be learned about its surface structure and composition from observing the small differences in that reflected light, even if it is very faint. Results are expected over the coming months as the scientists analyse the data being collected at this enigmatic body, 67P/C-G.



  • Robin Sherman says:

    Nicely explained and very illuminating. Thank You Emily.

  • This image from Voyager 1 shows something of the relative colors and brightness of the Earth and Moon :

  • Lucas Modena says:

    Amazing article. Thank you!

  • Kurt Niel says:

    Amazing article – some add values:
    Illuminance x albedo = reflected illuminance:
    Earth: 100,000 lm/m² x 0.31 = 31,000 lm/m²
    Moon: 100,000 lm/m² x 0.12 = 12,000 lm/m²
    67P: 10,000 lm/m² x 0.04 = 400 lm/m².

  • Stu says:

    Fantastic post, very informative, and another excellent example of the outstanding Outreach work ESA and the ROSETTA team are doing,

  • Andrew R Brown says:


    That is not from Voyager 1, that image was from the Jupiter bound Galieo spacecraft. If I remember correctly, the Moon was brightened about seven times, the scale is correct but the moon is much brighter than would be visible to the eye..

  • Cesare Guaita Milano Plantarium says:

    Thanks a lot for this interesting lesson !
    But a basic question remains: what’s the
    reason of the black color of the surface?
    In other words, what ‘s the main result of the spectroscopic inquiry of the surface material? In particular, what about the VIRTIS IR spectra in which there are rumors about strong peaks of organic materials? What are these organic materials?
    Are these materials the real source of the very black color? Thank a lot for a reply.

    • Robin Sherman says:

      The theory is that cosmic and solar radiation “cook” many of the organic molecules until they are reduced to carbon. If the “dirty snowball” analogy is correct, some of the “dirt” could be carbon as well.

      The phrase “dirty snowball” was a simplistic, media friendly sound bite chosen to represent what scientists thought a comet was like. It is unfortunate that this “best guess”, simple analogy, has become, from the public’s point of view, the only and precise view of what comets are . Probably because evidence from subsequent comet visits has not found enough evidence to prove or disprove this view and come up with a more accurate and descriptive short definition. I fear comets will always be described as “like a dirty snowball, BUT…….”.

      • Marco says:

        @Robin Sherman. There are several competing alternatives to the “modified dirty snowball” idea. I call the main ones – Wet comet theory (that comets have a liquid water core, among other things) , electric comet theory, which is the opposite – no water but dry rock and electromagnetic phenomena.
        Also my favourite i call living comet theory.
        The advantage with the modified dirty snowball idea, is that it fits in with observations of accretion disks in the birth of stars and how molecular clouds evolve.

    • Claudia says:

      Dear Cesare,
      As mentioned in the post, Rosetta scientists are using a variety of data to study the composition of the surface of 67P/C-G. Thorough analyses will be presented in the form of scientific papers in the coming future.

    • Cometstalker says:

      Tar is organic material and black as well. Organic does not mean created by life. It is just a bit semi complex to very complex chemistry containing carbon. Tholins are created on a lot of places without life.

  • logan says:

    Thinking of albedo. So, I was wrong, many ‘caves’ were just collapsed walls…

  • Cometstalker says:

    this shows how poor digital cameras perform compared to the human eye. If a person would be able to observe the comet from rosetta this person would clearly see the star-field haze of the milky-way, the stars and planets and of course the comet itself as the second brightest object on the local sky. The cameras on Rosetta have a dynamic range of 4000 to 16000 levels and the human eye has over 1 000 000 levels, the difference in pixel resolution is also gigantic. 4 mega pixels to 130 mega pixels. In sum to make an image with a camera that would represent what a human would be able to see in situ requires all the tricks in the image processing box and a lot of images. Then you still have a poor display to watch.

    • Cometstalker says:

      On top of this we also have an excellent image processing system that is doing the job instantly.

      • Jim J says:

        And we can make those wonderfully accurate eyes using unskilled labour!

  • Dave says:

    You may have written tongue I cheek bit which bit of the comet looks like a dirty snowball to you.
    I still can’t get past what looks like rock is a rock, especially still no ice found.
    Only a few days to the landing to find out, but will we be told – only ESA can tell us

    • Robin Sherman says:

      Well the “dirty” bit was explained above by Claudia, its the nature of the “snowball” underneath that dust, that is in doubt.

      Evidence that the comet is made of Rock.
      1.) In a few black and white photographs it looks like all there is, is dust and rocks like here on Earth, or the Moon and Mars.

      Thats it! Not a shred of scientific data or a single measurement to confirm that assertion. The Earth looks flat when you are standing on it, but we all accept that its a sphere. First principle of any scientific investigation, never trust what you observe until other non visual evidence is available to explain what you are actually looking at. Reality and perception are two different things..

      Evidence for the comet being largely made of volatile ices.
      1.) The comet originated in a part of the solar system where the predominant materials making up objects with known composition are gas, volatile ices and dust.

      2.) Previous measurements of other comets show large amounts of theses volatiles and dust in the tail of the comet. This material has come from the nucleus of the comet and so it is a reasonable assertion that comet nuclei CONTAIN volatile ices and dust.

      3.) Rosetta instruments have measured these volatile compounds, Water, Carbon Oxides, Ammonia and Methanol in the tail and coma of 67P. Thus its nucleus also contains these molecules.

      4.) At the temperatures recorded for the nucleus of 67P, these compounds, when present as part of a solid body in a vacuum, would be in the form of ice. These ices would sublimate to gases under these conditions to create the gas seen in the comet’s atmosphere or coma.

      5.) Flow rates for Water ice sublimating from the comet nucleus have been measured by the MIRO instrument on Rosetta.

      6.) The average density of the comet has been measured to be about 40% that of Water and incidentally, about 15% of your average Earth/Moon/Mars Rock.

      People can argue about measurement errors, but an 85% error? Conspiracy theories about ESA scientists and managers “fixing” the figures to maintain the current theory, are invoked. Why would you spend over a billion pounds to build a spacecraft to go and find out, if the objective was to maintain, what mission plans, proposals and media packs say is a theory full of unknowns and with insufficient measured data to reliably confirm?

      “It looks like Rock, so it is Rock” “I can’t see any ice, therefore there is no ice” are completely unscientific and irrational statements and yet they are used as the basis for justifying all manner of other assertions. The 67p nucleus may have some small proportion of Rock on its surface or inside it, but the comet is not made of Rock.

      A “rock” is a generic term for a solid object of certain shape, surface appearance and size, but this descriptive term does not determine what it is made of. Similarly, “pebble”, “stone” and “boulder” are generic terms for such solid objects of different sizes. They could be made of wood or polystyrene or plastic and then very well painted to look like rock, especially in a B&W photo taken from 10+Km away. Similar descriptive generic terms “puddle”, “pool”, pond”, “lake”, “sea” and “ocean” are used to describe different sized bodies of liquid. In our Earthly experience a “rock” is made of various combinations of minerals which varies hugely between different “rocks”. Our experience of using the same type of generic terms for bodies of liquid tells us however, that they may not be water, they could be alcohol, or liquid Nitrogen or Mercury or any number of other mixtures of chemicals. The rocks we see in these pictures are not Earthly and so to assume they are made of the same minerals is, to be polite, unwise. Especially in the light of the experience of the objects that “looked like” rocks and stones on Titan, but which were in fact when measured with a spectrometer, made of Water ice.

      As a scientist myself, a chemical analyst to be precise, I cannot afford to assume anything about any material, gas, liquid or solid. My safety depends upon it. Rosetta and Philae’s safety depends upon the density calculation. That safety is not going to be put in jeopardy to save the blushes of a few no name scientists. It is to be hoped, as you say, the results from Rosetta and Philae conclusively prove what 67P is made of. Unlikely, as those who choose to will claim, its just one sample, or the electrical environment of the comet messed up the instruments, or its all made up CGI in a studio in Paris, whatever.

      Me, I’ll weigh what evidence is presented and make up my own mind. So far all I can be sure of is, its a very, very dark object, with unreliable visual evidence of significant amounts of volcanism and erosion on its surface. It has a density that is significantly less than expected by current inadequate theories and that it is made of a mixture of small amounts of Rock and organic compounds, lots of dust and volatile ices in an, as yet to be explained, very porous structure. In short we the public are stuck with the phrase “very dirty snowball(?)” as the sum total of the knowledge imparted to us so far.

      • Robin Sherman says:

        I should just add it is Claudia’s piece above that has added the “very” to that knowledge. More informative explanations like this of the topics discussed here please.

        The (?) is the reason for and purpose of, the Rosetta mission and for that I applaud the scientific and educational ethos of ESA and the mission team.

      • logan says:

        A few dozen photos and a handful bits of data. Up to today.


        (Nowadays, grateful for them)

        • logan says:

          Still not quite sure, but our safety as ‘invited’ public is not jeopardized by our tantalizing comments. And that is our advantage too. We can easily express the more wild ideas. We can diverge and dissent. And this is good 🙂 If some of the People on Board wants to distract a moment, (s)he can drop by, smile, forget, reboot.

      • logan says:

        A trashy, dirty, icy dustball? 😉

  • Mark Zambelli says:

    Brilliant article Claudia, very nice.
    (The optical illusion explanation has a typo… ‘lighter’ appears twice; the first instance should read the squares in shadow are ‘darker’) 😉

    It’s amazing just how dark these objects really are. I’m an amateur astronomer who loves astrophotography so it was very insightful to read your blog, bravo. The ‘anchoring effect’ you mentioned is so powerful it’s astonishing the full moon is so dazzling and I always get looked at strangely when mentioning how low the albedo really is. Rosetta is doing a marvelous job.

    • Mark Zambelli says:

      Moderators…OMG please delete that mention of any typo, I apologise… I’m an idiot, sorry.

  • logan says:

    Hi Claudia. Really enjoyed your article. Could go directly to textbooks. Not the boring, antique, committee approved material actually there.

  • Jim J says:

    let me gently ask in the spirit of scientific observation, what evidence is available on the early history of 67-P that records or otherwise positively confirms where in the solar system it was created? Do you know for sure how it was created? People would like to hear that if it’s confirmed observation.

  • Stan Zorin says:

    The ‘checkershadow illusion’ example is wrong. The square A is darker than the square B

  • Duke says:

    I am guessing that the body of the comet is pure iron. If not, it will be a pure oxide of one metal, perhaps silicon dioxide.

  • AlfonsoFR says:

    Wanna see colorized pics of the comet? Here get these

  • Michel Rouzic says:

    The comparison image which shows the Moon as being very dark was made by someone who obviously has no understanding of how gamma compression works and just assumed that sRGB value are linear.

    Or to put it in simpler terms, the Moon is obviously not that dark, it’s just that the person who made that picture has no understanding of how computers handle colour.

    12% albedo gives you a sRGB value of 97, not 31, and a 5% albedo gives you a sRGB value of 63 and not 13.

Comments are closed.