Mars Rover Picture Analysis Discussion

Firstly, thankyou Barry for those comments. I too have been wondering about the panoramas with the same name. The second is obviously a further processed version of the 'cherry kool-aid' image. It seems a lot of the early shots are rapidly, largely automatically mixed and stitched, and released for the press as quick as possible. While later versions have had been calibrated more thoroughly. Dr Bell suggested it might take months to fully calibrate and release all images.

As far as the use of the L2 filter, as far as opportunity, the initial few images from the rovers are used to try and choose areas of interest. To decide where to send the rover after it rolls off the lander. I would suggest the L256 images are preferable for the larger amount of data they can supply. To be honest I think I have to shore up my knowledge in the field of geology to understand this further. I do know that most of the Rovers tools rely heavily on IR spectroscopy to learn the composition of the surrounds.

Another area I'm looking into is the near-IR reflectivity of all man-made pigments. As a good deal of the colors on the rover show an increase in the near-IR region. It may be possible that this is a common thing to man-made pigments.

But for now, its Australia Day down here so I'm going to grab an esky and go to the beach to wait for the fireworks. Will get back to you in a day or two.

marsrovers.jpl.nasa.gov...

marsrovers.jpl.nasa.gov...

look at these its just like the black dot moved from were it was. take the crater and align them so they are upon each other so it looks like one but there are to pics. you will see that the black dot moved away.
-----

doing this minds you how easy it is to think they are doing something to the pics cause they are the desent pics and its propebly the shadow of the desenging craft.

[Edited on 26-1-2004 by MarkLuitzen]

EDIT: Pictures killing layout, converted to links.

[Edited on 27-1-2004 by Kano]

Don't get me wrong... I think that the L2 lens is highly useful in the data-gathering aspects of the Rovers, particularly when used in conjuction with R2 for binocular pictures (3-d perspective). I'm sure that the L2 lens also provides plenty of useful information to the geologists when trying to find interesting rocks.

The real issue that I have is when they try to move from target selection to presenting "color mosaics". As near as I can tell, no significant additional SCIENTIFIC or navigational information comes from mixing L2 with L5 and L6, above and beyond what any one of those single-filter pictures would have given the mission team.

The primary function of a "color" mosaic (as far as I can tell) is public relations. It makes prettier pictures than the black-and-white of a single-filter exposure.

And make no mistake, the primary interest level of the MASSIVE downloading being done (of the data being made available) is for color pictures destined for "human consumption"... desktop wallpaper, flights of fantasy, and generally making folks feel good about their investment. It makes the public interested and enthusiastic, which is good for an agency that is funded with tax dollars.

It's fine to take L2 frames, and I'm happy that NASA is doing so. Where I'm bothered is with the logical leap that, since they have the L2 frame, they can just substitute that along with L5 and L6 to create a "color" picture which does NOT represent something close to the human / computer color range.

The January 19th pictures of Sushi and Sashimi show just how well the pictures come out when they use the far more appropriate L4 lens to make "approximate true color" images.

That means that NASA has a chance to show people something extremely close to "the truth" when composing color pictures to show people "what it looks like on Mars". All they need to do is take the L4 frames IN ADDITION to the L2 frames.

Doing so lets geologists see the data in an L2-L5-L6 format if they prefer that for some reason (they don't need the NASA website or press releases for that... they can produce those easily in-house I'm sure), and the rest of the public can get L4-L5-L6 images which come really close to "true color"... and there are no more ongoing and compounding problems with credibility.

I understand the argument for wanting to put something out on the web "as quickly as possible", but please keep in mind that these web shots are NOT the primary mechanism for target selection in the future (as far as I know). I'm sure that the principal investigators and their teams are looking at the raw data, not surfing the website and press releases.

Press releases are for PUBLIC RELATIONS, pure and simply. That means that the color results should be based on data that is compatible with web surfers... miscellaneous human beings viewing the pretty color pictures on their computers using the Internet.

Parties interested in doing their own ad-hoc analysis can still download all of the raw data and perform their own analysis (subject to the minor errors introduced by reformatting the data as JPG files instead of releasing the lossless images, of course).

I just think that it is woefully short-sighted of NASA to neglect to take an extra shot at each landscape position using the L4 filter... they are constructing the pretty color pictures for PR purposes anyway.

So why do they sabotage their own PR efforts by putting out inaccurate (for humans) color pictures and then have to explain later why the real (human) colors are so different? Why not just do it right to begin with?

Their existing strategy is fueling conspiracy speculation FAR more (IMO) than if they would just "do it right" from the beginning. Spend a few extra minutes, and get it right the first time.

Take the L4 shots along with the L2, L5 and L6.

Putting out the L4-L5-L6 shots first puts a damper on the conspiratorial whisperings, because you don't have to deal with the accusations of a "changing story", which gets otherwise-accepting people scratching their heads and questioning (in many cases, for the first time) if there is a coverup.

I know that I, for one, wouldn't have batted an eye about the pictures if the color panoramas had been done in L4-L5-L6 initially, including the calibration tool. The presence of that tool would have allowed any reasonable person to verify for themselves that the L4-L5-L6 mix was pretty close to "right on", and I'm sure the vast majority of them would have dismissed outright any of the more loony conspiratorial excursions.

Instead, we're left with a situation where I see that calibration tool, I get (along with almost everyone else that I've pointed it out to) a visceral reaction that I'm being deceived.

That's horrifically bad PR. Instead of the relatively small "usual suspects" list of conspiracy fanatics raising the typical fuss that can ALWAYS be expected, we are left with mainstream news sources like the BBC pointing out the color discrepancies.

People pay a LOT more attention to something like the BBC news than they do to the typical conspiracist.

Even with the relatively prosaic explanation that Kano has uncovered for the MECHANICS of why it looks that way... we're still left with the contentions made all along by team representatives that we would see Mars "in its true colors", and then procedures that more-or-less guarantee that we can't see the truth because of the flawed methodology in data gathering.

The red channel on L2 is contaminated when trying to make human-friendly pictures. We are then left with reassurances that it "probably" looks very close to that... only to find other pictures taken and balanced with the human-friendly L4-L5-L6 combo that show that the terrain clearly doesn't look like that.

That's not a recipe for recovering lost confidence, IMO. That's a recipe for continuing to fuel speculation about why they won't show the rest of the landscape as a human would see it.

(sorry, accidental double-post)

[Edited on 1-26-2004 by BarryKearns]

I concur that it was somewhat of a miscalculation to release the L256 pictures as the first few images from the rovers. The ideal (from a PR point of view) thing to do would things in the order Barry mentioned, release L456 images first. Then make a point of mentioning that most images would from then on be a reconstruction using a shifted red channel.

I also suspect that the reasoning for this is NASA trying to be as mission-focused as possible. Perhaps not listening to the PR people as much. But the counter-argument to this obviously is it would look just as bad to use up too much bandwidth sending extraneous L4 plates. Especially to the people in charge of the funding and planning of future missions. As far as the mission goes, the primary objective must be results. If Spirit had have died for real the other day. It wouldn't look too good for NASA if they had spent too much time sending L4 plates.

As well as this, of course, is the fact that the conspiracy mind knows no bounds. Even if the first few images were in L456 for the public, then later the images became L256. We can be sure that people will cry foul anyway. No matter that they have been told it was going to happen.

So yes, I do think it may have been better to release L456 images first. But maybe we spend too much time browsing these types of forums

marsrovers.jpl.nasa.gov...

and what is this in the middle of this picture

found on marsrovers.jpl.nasa.gov

sol 2 raw images.



[Edited on 27-1-2004 by MarkLuitzen]

Originally posted by Kano
As well as this, of course, is the fact that the conspiracy mind knows no bounds. Even if the first few images were in L456 for the public, then later the images became L256. We can be sure that people will cry foul anyway. No matter that they have been told it was going to happen.

So yes, I do think it may have been better to release L456 images first. But maybe we spend too much time browsing these types of forums

You're welcome to leave. Let's not downplay the significance of the L4 filter simply because 'inquiring minds' will 'never be satisfied'.

BarryKearns, thank you very much for your contribution to this thread. You've cleared several things up for me in a way that I can easily understand. Great job!


[Edited on 1/27/2004 by Bangin]

Thanks, Bangin.

Kano, re:

But the counter-argument to this obviously is it would look just as bad to use up too much bandwidth sending extraneous L4 plates. Especially to the people in charge of the funding and planning of future missions. As far as the mission goes, the primary objective must be results. If Spirit had have died for real the other day. It wouldn't look too good for NASA if they had spent too much time sending L4 plates.

I really don't think that the bandwidth/time argument carries a lot of weight in this case. The goal of human-friendly pictures would have been accomplished quite well if they had simply chosen to include an EDN frame using L4. The typical mosaic shot was composed of (at a minimum) an L2 EFF frame, an L5 EDN and an L6 EDN. The EDN frames are only 25% of the data size of the full EFF frames, due to downsampling to 512x512 from 1024x1024.

That means that worst case, including an L4 EDN frame with each of the color mosaic frames would have increased the data size by only about 16% or so... far less when you consider all of the frames being taken for purposes other than mosaics, and all the frames being taken in a wide variety of other filter formats.

Taken as a whole, the L4 frames would have been a quite-small fraction of all transmitted data.

In fact, taking the L4 as EDN would make the combination of the mosaic images much EASIER, since the resolution would be identical on L4, L5 and L6... no resizing necessary, which means fewer artifacts introduced! The L4-L5-L6 in EDN might actually appear slightly SHARPER than the L2-L5-L6.

I understand the "mission" argument, but I think that the inclusion of the L5 and L6 EDN frames shows that the team is making reasonable compromises on data payload in order to meet the additional goals of communicating effectively with the public... all I'm pointing out is that they took only two of the necessary three steps to make that goal a reality.

The actual exposure times are quite small, if you look at the time stamps from consecutive filter shots... it looks like it's typically about 32 seconds to switch filters and take another frame.

So it's not time consuming, it doesn't add substantially to the data payload, and it makes the communication process (PR efforts) vastly more effective... yet they are choosing not to do it.

Not a good move, IMO... not a good move at all.

(and before anyone points out the "number of files" argument regarding the OS crash... I'm pretty sure that the many thousands of in-flight data files that they left littering the Flash RAM had much more of an effect than the addition of the relatively small number of L4 frames would have.)

Come now Bangin, you know as well I do that there are a large percentage of people who choose to believe as they choose no matter what the arguments for or against. However, that is another matter.

I'm not downplaying the significance of the L4 filter, but it is apparent that for the majority of purpose (especially ones directly related to the mission) the L2 will suffice, and is even preferable. I concur that it would have been nice to release at least a few L456 off the bat to keep the public smiling. But my guess would be the PR people at NASA lost that argument. To send everything with L2 L4 L5 L6, even though it is only maybe 10-15% additional data from the PanCam. Would still mean losing 10% or so of the total PanCam data that would otherwise have been transmitted. As for mission purposes L4 is largely covered by the L2. Thats still a fair percentage of data to discard after spending so much time and effort optimizing the PanCam and systems.

So I agree a few shots is a good idea, but still think that a mission orientated stance is perfectly understandable. It looks like its just a case of trying to spin too many plates at once, focusing on the mission while still getting workable images out to the public ASAP. It will be interesting to see what the final images come out like in a few months.

Of course theres also the possibility that the increased attention might strengthen the PR-depts case somewhat, and allow more and more L456 images to be transmitted. After all, they have already started to make a note of the fact that the images are approximate true color images.

Originally posted by MarkLuitzen
marsrovers.jpl.nasa.gov...

and what is this in the middle of this picture

found on marsrovers.jpl.nasa.gov

sol 2 raw images.



[Edited on 27-1-2004 by MarkLuitzen]

Looks like some guy whose running like a bat out of Hell.... nosy journalist who has found the site?

Blessings,
Mikromarius

hmmm... there seem to be several different frames with areas of unusually high reflectivity across all frequencies taken.

Another good example: marsrovers.jpl.nasa.gov...

Could this be some kind of an issue with the auto-leveling process taking a picture that is almost uniformly dark (that soil background) and one piece of uniformly light gray colored rock, and amping up the exposure to get it to a pixel value of 255 (giving a brilliant white appearance)?

Originally posted by BarryKearns
Another good example: marsrovers.jpl.nasa.gov...

It's maybe the same dude, only in this picture he is shot dead by MiB. Darn! And he was sooo close!

Blessings,
Mikromarius

Originally posted by BarryKearns
hmmm... there seem to be several different frames with areas of unusually high reflectivity across all frequencies taken.

Another good example: marsrovers.jpl.nasa.gov...

Could this be some kind of an issue with the auto-leveling process taking a picture that is almost uniformly dark (that soil background) and one piece of uniformly light gray colored rock, and amping up the exposure to get it to a pixel value of 255 (giving a brilliant white appearance)?

Yes, it appears that way.

Although it isn't necessarily gray. Just the brightest parts of all channels. So it is just the lightest part of the image.

[Edited on 28-1-2004 by Kano]

From Dr. Adler

quote:
--------------------------------------------------------------------------------
Yes it does send back a bunch of information about the image, like the exposure time, camera pointing, temperature, etc. However as far as I know, that data is not available online. Once this data is archived in a few months, all of that data will be included and documented. All of the mission data will be available at the cost of duplication.

--------------------------------------------------------------------------------

Some of this information is indeed available now, for those images which have been included in the Maestro datasets. I invite people to download and use maestro mars.telascience.org... while paying close attention to the Maestro\Wits-DB\MER\Spirit\common\state\(timestamp).msml files. They contain such information as "INSTRUMENT_STATE_PARMS.INSTRUMENT_TEMPERATURE" and "INSTRUMENT_STATE_PARMS.EXPOSURE_DURATION". This, combined with information from such papers as europa.la.asu.edu:8585... and robotics.jpl.nasa.gov... should allow more "true" color images to be created.

Just from what images have been published already with Maestro, it appears that when a L2 R2 L5 L6 set is taken (not including the other filters), they have relative exposure times of ~ 1 (L2) : 4.4 (L5) : 6.5 (L6). Add in the information regarding the relative sensitivities of the Pan Cam CCD (the above bell_2003.pdf paper) and you have much more realisitic idea of how much light is actually being thrown around in different wavelengths, at least for those particularily that have been measured.

[Edited on 29-1-2004 by slinted]

EDIT: Fixed links.

[Edited on 31-1-2004 by Kano]

Excellent find, thankyou slinter.

Also, from the latest press release:



Larger image here:
marsrovers.jpl.nasa.gov...

Ground sky and sundial in one frame. (Sky in sundial mirrors).

Originally posted by Kano

Ground sky and sundial in one frame. (Sky in sundial mirrors).

Definitely an encouraging development! I read on the Space.Com pages that someone had spoken with Dr. Bell, and that they were planning on releasing L4-L5-L6 panoramas in the next few days, so the good news just keeps on coming...

I'm a little disappointed that they would construct and release something like this without also releasing the associated raw data, though... they obviously must have it, so I can't see what the hangup would be in releasing it.

Instead of the heavily cropped-down sundial pic for the composite, I'd like to put the full frames together myself and see what the balances look like.

All in all, a pretty happy day, I'd say!

I don't think the process to get the raw images online is automated. I'm hanging out for the raw data also.

I also suggest anyone who hasn't already grab a copy of the 2 .pdf files linked by slinted. These are the relevant articles submitted to the JGR(Planets) by Dr. Bell and associates. They contain a pile of interesting info about all the Cameras on the Rovers, specifically the PanCam.

Originally posted by Kano
I don't think the process to get the raw images online is automated. I'm hanging out for the raw data also.

I also suggest anyone who hasn't already grab a copy of the 2 .pdf files linked by slinted. These are the relevant articles submitted to the JGR(Planets) by Dr. Bell and associates. They contain a pile of interesting info about all the Cameras on the Rovers, specifically the PanCam.

Yes, the .PDF files are definitely great resources... but the first of them references at least what Dr. Bell was PLANNING on doing with the raw images:

"The primary Pancam standard data product will be Reduced Data Record (RDR) image files in PDS format and calibrated to radiance. These EDRs and RDRs will be archived with the PDS and released to the community on the schedule outlined in Arvidson (2003). In addition, EDRs will be converted to JPEGs and streamed onto the World Wide Web in real time as the downlink is received. "

(italics mine)

According to that, the raw images should hit the Web as soon as they reach the downlink, not multiple days later (if at all).

Is there anything that indicates why they are deviating from that plan?

[Edited on 1-31-2004 by BarryKearns]

Is there anything that indicates why they are deviating from that plan?

Negative, I think the entire time the images have been updated manually, as there have been little comments going along with the release of each days images. Also the layout has been changing as the missions continue.

[Edited on 31-1-2004 by Kano]

Originally posted by Kano
Negative, I think the entire time the images have been updated manually, as there have been little comments going along with the release of each days images. Also the layout has been changing as the missions continue. Not 100% sure though.

Actually, I don't think the two concepts are that incompatible. It would certainly be possible that the raw images could have been streamed into the necessary file directories automatically, and what we are seeing as a delay could simply be the updating of the associated web-page front end that gives easier surfing access to the images.

The comments and redesign of the front end wouldn't be incompatible with having the raw JPEGs dumped in there in real-time... of course it does us little good if they don't have either some kind of basic directory-dump page generator or allow us to directly browse the directories in question (which they don't).