MEX VMC - Back on, and online! Options

[remcook]

yes, I think 80 km is a normal altitude for these mesospheric clouds, but I think these are normally not that bright, especially when looking down on it.
For instance, the CO2 clouds reported here are not visible:
http://www.esa.int/esaCP/SEMC4JZ7QQE_index_0.html
What latitude is this roughly? Apparently, mesospheric clouds are common around the equator:
http://www.agu.org/pubs/crossref/2007/2006JE002805.shtml
This article came up here before and shows similar clouds+shadows:
http://www.esa.int/esaCP/SEM1DV3MDAF_index_0.html

[ormstont]

Wow...you beat me to posting the link I've just been chatting again with the Omega team (the ones that wrote that paper that your last link is based on)...if anyone knows Mars clouds it's them! We had a quick look at surrounding features and worked out the VMC cloud would be around 18 degrees South and 162 degrees East, which would put it close to the equator but still away from a common region of these clouds (based on Omega observations).

More generally all the MEX instrument science teams responded very warmly to the results produced by yourselves and other VMC contributors and may even see about giving us more VMC observations as a valuable contribution to Mars Express science. We'll put it all up on the VMC website though as always, as soon as it's down on the ground (next dump should be Friday/Saturday, so look for more new pics at the start of next week!).

[elakdawalla]

I'd like to include one of the VMC crescent images in my "Year in Pictures" roundup for The Planetary Report but am so far not happy with what I have been able to produce. Attached is the result of my fiddling with the crescent image from the "Best of 2007" feature. But I need an image from 2008 for this roundup. There are a lot of crescent shots in the archives from the spring and summer of 2008, though most of them appear to be saturated. I'm going to keep fiddling with the data -- these images are really fun -- but I know there are lots of people here who are better than I am at making pretty Mars color images and am hoping that one of you will be able to come up with something more spectacular than I can.

--Emily

Attached thumbnail(s)

 

[ugordan]

All 2008 crescents indeed appear to be severely overexposed. The closest thing to a non-overexposed crescent view are images from June and July, but they're already at a fairly fat phase. Here's one of the later ones:

It's looking down on the southern hemisphere with a high cloud peeking over the terminator.

If you're willing to put up with overexposed high phase shots and have an image that doesn't really scream "this is Mars", might I suggest one from that sequence showing clouds at the terminator that was brought up here, from Feb 14th.
FWIW, either Mars isn't terribly reddish at those high phase angles or the VMC data is very noisy, or both. Getting anything out of those thin crescents except a greenish hue seems like an exercise in futility.

[ormstont]

So, looks like our other MEX instrument teams are working hard to find out more about our strange cloud - looks like we could be in for more too - they said that this is the season we'd be expecting these clouds on Mars (although the VMC cloud is larger and more pronounced than anything seen before!). Will keep you posted when I hear more!

For the crescent images - sorry that they're all really saturated - those problems were one of the reasons I've moved to doing the 4 exposure step cycle (even though it cuts down the number per observation of properly exposed images) you see on the current images. The nice thing about the saturated images are that we see the features in the night side, over the terminator (I know...excuses, excuses!!). I promise to try harder at the next crescent season! Of course - any suggestions on how best to set the exposures or timing of images let me know - you're the operators of this camera!

[ugordan]

I'd be interested in knowing what the difficulties are in adjusting exposure for a given observation. Is it the uncertainty of how bright Mars will be at certain phase angles?

While we're at the subject of exposures, knowing how long each raw image was exposed would be of additional help in processing.

An unrelated observation - it seems to me that if only we had a flatfield frame for the VMC we'd be able to get much better images out of it. A lot of the noise seems static on the detector.

[ormstont]

Well, part of the difficulty is that we have no way of predicting how bright Mars will be in a given location at a given phase angle. I'm sure one exists but our aim is always to allow the public to guide VMC activities - as I say, you are the operators of this instrument. If anyone comes up with a way of predicting it then I'd be more than happy to implement it!! I implemented the 4 step exposure setting as a workaround for this - guaranteeing whatever the observation we should get some well-exposed shots.

That's a good point about knowing the exposure settings - I'll see how we can work that into the routine flow of information and processing of each image. For a guideline though, the first image of every observation is always 14ms. For the rest of the observations I'd have to look for you on Monday at work, will do that and get back to you then.

For the flat field - I'm vaguely aware of what the term means - would it be a long exposure image with the spacecraft pointed at space? If you let me know roughly what I need to command MEX/VMC to do then I'll try and schedule a slot when we can do it. I know I use this as a mantra...but it's everyone's instrument, you let us know what to do!

[ugordan]

would it be a long exposure image with the spacecraft pointed at space?

That would be a dark frame, used for figuring out how the background signal increases with exposure (varies with temperature, hot pixels, etc). Part of wanting the exposure setting comes from the desire to have a rough model for dark background - two separate exposure images of a same target will produce different colors for different amounts of background signal. For example, a short exposure with VMS typically shifts overall colors towards blue as the background (grayish) overwhelms the Mars signal.

A flat-field is usually produced by imaging a sufficiently bland target so any noise in the image is variations in pixel sensitivity, crud in the optics, vignetting etc. This can be divided out from subsequent images to provide better S/N ratio. I don't expect we can get a look at a bland target now that we're at Mars, that sort of thing could have been done back on Earth, but then again this is not a scientific camera so most likely wasn't "calibrated" this way.

[Tman]

A flat-field is usually produced by imaging a sufficiently bland target so any noise in the image is variations in pixel sensitivity, crud in the optics, vignetting etc. This can be divided out from subsequent images to provide better S/N ratio.

For example, the Mars Rover's lenses have also a couple of "noises" by now, beside their vignetting.
This image shows clouds on sol 1647 and this image from sol 1645 only the dirt on the lens. As a "fieldwork" flatfield you can use the inverted sol 1645 image for some clearing.

[Paolo Amoroso]

I don't expect we can get a look at a bland target now that we're at Mars, that sort of thing could have been done back on Earth, but then again this is not a scientific camera so most likely wasn't "calibrated" this way.

Would this Iris technique for extracting a flat field from science images, or something similar, be useful with VMC images?


Paolo Amoroso

[ormstont]

Ok, get what you're going for there - the only suggestion I could possibly make for an in-flight test would be an exposure of Mars at low altitude that's long enough to blur out any surface detail (although I guess that would probably saturate the detector too).

I'll see at work tomorrow if any tests were made pre-launch although I think most of the dirt/objects/scratches(?) we see on the optics came after Beagle 2 ejection. VMC is right underneath the Beagle ejection mechanism so there might have been outgassing or debris from the pyro bolt firings that separated Beagle (some small objects can be seen moving away from the spacecraft in the Beagle separation images).

Based on this possibly the technique that Paolo mentions with this software could be helpful? Would certainly be great if we could extract the artefacts from each image.

Finally, you mention what we really need is a flat-field. Would a dark frame be helpful too? I've never tried VMC pointed away from Mars, so would be interested to see what happens (more out of engineering curiousity)!

[ugordan]

I don't personally think a dark frame would be all too useful for purposes of processing images alone - the detector is 8 bit and the majority of the noise is flatfield related noise. That said, I wonder how long an exposure can you get wih the VMC? It's a wide angle imager so we can forget about imaging individual stars probably, I wonder if it'd be possible to capture the Milky Way glow at least? It would probably depend on how fast dark current builds up with exposure.

[jekbradbury]

I am trying to organize the images by position in the orbit, but the quoted orbital period(s) on the Mars Express website (there are two different numbers on different parts of the site) don't seem to match the images. Taking the time since the beginning of 2008 modulo the orbital period of 24180 seconds should put the partial disc images near each other, but doesn't. What is the actual orbital period? Alternatively, would the SPICE kernels have this data?

[ormstont]

That said, I wonder how long an exposure can you get wih the VMC?

Ok....first off - the VMC exposure range is pretty wide. There are two exposure ranges supported -

• 0.4 ms to 191.6 ms in increments of 0.8 ms
• 200 ms to 95800 ms in increments of 400 ms

The widest range tested was in our first two observations in 2007, the first where we almost missed the planet here: 2007-01-28 Observation, and the second where we got our first disc image here: 2007-02-26 Observation. I'll look up the exposure ranges used here for you, but I know we only just touched the longer range.

For a long exposure test pointed at deep space I might see what I can schedule as I'd be interested to see how VMC performs in those conditions too, I'll keep you posted on my progress.

I am trying to organize the images by position in the orbit

Not sure exactly what you're trying to do here, but please let me know a bit more and I'll see how we can help. If you mean what point in the orbit, almost all VMC observations were taken just after apocentre. If you mean ground location it's a bit more tricky as the MEX orbit precesses in Argument of Periapsis to allow our pericentre to not only be over different longitudes but also different latitudes (although longitude obviously varies rapidly with Mars spinning, the latitude precession is much slower, over the course of a year).

What is the actual orbital period? Alternatively, would the SPICE kernels have this data?

Ok...the MEX orbital period has changed throughout the mission. SPICE kernels will give you a full, complete and accurate definition of the MEX orbit, but there are some shortcuts you can take. If you go to the MEX SPICE server here: ftp://ssols01.esac.esa.int/pub/data/SPICE/MEX/kernels/ and go into the orbnum directory you'll find a .ORB file that lists all the MEX orbits. The column OP-Event UTC APO gives the apocentre time for each orbit - so the period is the duration between any two adjacent values in this column. For even more detail on the MEX orbit you can use the EVTM_ files at the auxilliary data server here: ftp://ssols01.esac.esa.int/pub/data/ESOC/MEX/ (Just take the highest version number file for the latest).

[jekbradbury]

That was exactly what I was looking for. Taking into account the actual average period for the time period of these images, they are correctly placed in the orbit. I had hoped there would be some pictures at every point in the orbit, but it seems the images are in two sections: a ~1hr period near apocenter, and another ~1hr period near pericenter. My goal was a simple animation of what VMC sees on a typical orbit, not taking into account phase or ground location, just trying to show the varying size and brightness of the disk. I'm assuming operations (downlink etc.) and other instruments prevent this from being done, but one VMC image every ~5 minutes for one full orbit would make for one amazing animation.