Oppy sol 1311 colorized MI panorama:



with links to location and original panorama -- which has links to 3D pairs.

An experimental MI dust correction was applied to the images before autostitching the panorama.

Brinesplat vugs, happens all the time in Barsoom...

Hello,

In late for many images, but that's all that I can do now.



I have trying to eliminate dust darkening with a flat picture I've build from a sky left pancam observation. I give you the flat-dust-corrector (not very perfect on all the images, optimised for L2 pics). I'm working on obtain better results (this flat do not correct over-exposure on the left pictures due to a dark part of the image). Work with a simple addition mode.

Nice tool, Ant!
I tried it and works, however the sum gives a reduced contrast on the right portion (as foreseable). See below for example:

Perhaps, a multiplication flat field would behave better (I'm not sure, however)... what do you think?

I read Skyelabs link and it seemed to be saying that vugs in carbonates were often formed by solution by water. This made me think that the vugs in the white band of Victoria crater might be caused by solution by near surface ground water shortly before Victoria formed.

Perhaps there is now evidence of water at merridiani both at say 3.5 billion years ago when the evaporites formed and at say 2 billion years ago when Victoria formed. Perhaps this is part of the same story as the dessication cracks seen at the bottom of Endurance crater.

PS I appologise for multiple posts. It is difficult to post using my mobile phone.

You need to do a bit of both (and the amount is dependent on the amount of stretching).

James

to flatfield compensate an image you just divide the data with a normalized flatfield image. multiplication or addition is just not right in my book...

/mattias

OR you can multiply with the inverse of the normalized flatfield.
The reason it's not as simple as that with jpegged raws is they are histogram stretched and so is the flatfield effect. You'd have to something like stretch the flatfield the same way before dividing to remove the effects, anything else is bound to leave residual noise.

I've entertained an idea that if you had the proper flatfield (by proper I mean it really minimizes high frequency noise and other large scale nonlinearities, but principally noise) and if you had a linearly stretched raw image, in theory you should be able to work out the rough inverse of the histogram stretch by dividing with the flatfield while modifiying lowest and highest DN values until the output residual high frequency noise is minimized. It would then give you the rough initial min. and max. brightness levels before the stretch. Of course, one has to consider what metric (Fourier filtering?) do you use for determining when the residual flatfield noise is the smallest (as opposed to actual high spatial frequency real details). There's also the fact MER images are mostly LUT encoded IIRC, add in JPEG artifacts and this doesn't realistically pan out. Still, it's a thought. Might be only worth it for sky shots which are pretty bland anyway...

Why aren't the image planners just using more overlap and then just cutting away the bad parts? Will take more images to get the same pan, but it will look good

Agreed, more or less. Ignoring crystallography, the vugs might represent frost-leached or heavily-hydrated former chloride crystals, such as antarcticite or hydrohalite (both unstable much above the freezing temperature of water), or they might instead represent so-called meridianiite:
http://www.lpi.usra.edu/meetings/7thmars2007/pdf/3124.pdf
a heavily hydrated magnesium sulfate that is likewise unstable much above freezing, or they might even represent tiny ice crystals that later sublimed. Very difficult to say at this point without more data. Their extremely fine grain size and finely dispersed nature suggests to me that they probably are not normal evaporite crystals, formed by direct crystallization from either liquid surface water or liquid ground water, and likewise that they were never soaked in such warm liquids. They could have formed diagenetically, or immediately after deposition of the originally damp rocks in which they are found. As always, feel free to disagree.

BTW, if you look up the MgSO4-H2O phase diagram in that meridianite reference, you should note that the freezing point depression of magnesium sulfate, the most soluble common sulfate, is only about -4 degrees C, in contrast to common chlorides that (in combination with ice) can still form liquid brines at temperatures up to nearly 50 degrees colder. That is why we feel that very cold subsurface brines on Mars, if any, should be strongly enriched in chlorides instead of sulfates (contrary to what is commonly inferred from the sulfate-rich surface mineralogy). The sulfates might be trapped at or near the surface of Mars simply because they cannot easily be frost leached, except at comparatively high temperatures. Again, feel free to disagree.

-- HDP Don

Meridianiite! Thanks for pointing that out. After a little Googling I see that the earthly occurance of the mineral seems to have been discovered by the same Canadian guy who created the stuff from epsom salt solutions in his garage in the winter, and then went out to find it in the field. I recall reading about the garage experiment last year, but hadn't heard about the field discovery. It's a very cool story. I am no mineralogist. Is there any reason to suspect its crystals would fit into the observed vugs?

As for the "vugs," I'll agree that some of the things we are seeing in the recent MIs resemble vugs we saw back at El Capitan in Eagle Crater and other rocks right up to Endurance. But these seem noticeably smaller and less well defined than some of those observed early in the mission.

Another thing I can't help but notice is that I really can't identify any original grain outlines in the new MIs. I understand that we are looking down onto the bedding rather than at a cross section of it, but I can't help but wonder if this rock has been significantly altered. I don't know how much the MIs have been degraded by recent dust accumulations on the lens, so that makes interpreting the raws difficult. I'd like to see some MIs of a vertical cross section next.

Oppy sol 1294 L257 2x3 panorama:



Lots of synthesized missing data. I applied a variation of the Ant103 dust mask as a division by the inverted mask plus an offset.

I am ready to give up developing a simple "ACME dust correction" (beep, beep.)

Please, please JPL, apply the dust correction before creating the "raw" JPG images!

I'm going to wait for the missing data for that one i think.

Here is the view on the other side on the same sol. L7 only at this stage, I'll drop in the colour when it arrives.



James

I too mad(e) a stitch of upper frames only, using Ant103 "flat field" and lot of work attempting to adjust contrast differences (not completely succesfull!):

NB: artificial sky

sounds pretty cool but even in really great conditions you would probably have an errormargin that is way bigger than just guessing...

another approach would be to use information from the earlier released raw images.

One could probably use the time of day (sun elevation) to estimate the DNs expected for known albedo targets in an image. the dusty sand seems to be pretty much the same color in every picture so far. so you only have to find one patch in shadow and one in direct sun and you would be pretty much golden...

/M