A City Under Cydonia?

December 14, 2002
Posted by: Ky 

On August 29, Richard Hoagland of The Enterpise Mission released a paper describing the results of their month-long analysis of an infra-red (IR) image received by the Mars Odyssey space probe of the Cydonia region of Mars and released to the public by Arizona State University (ASU) on July 24. Simultaneously, Hoagland made an appearance on the Coast to Coast AM program with Art Bell to publicize his findings. During the appearance, Hoagland alleged that a second image was passed to The Enterprise Mission (TEM) associate Keith Laney on July 25 which differed substantially from the original image available for download from the ASU website. Hoagland went on to claim that this image showed a city buried under ice at Cydonia. The second image was alleged to have come into Laney's possession by ASU staff specifically redirecting him to the image showing the claimed city.

Much controversy and debate subsequently ensued as to the legitimacy of the second image which came into TEM's hands. While TEM was adamant that the image that they obtained represented genuine data (claiming that the public version represented sanitized data with the alleged city removed), others suspected that the TEM version was a fake.

This article is part one of a two part article covering flaws in the story in the early days, with image comparisons to highlight certain aspects. Part two will conclude with a precise explanation as to how the "city", and the public, were hoaxed.
In the initial TEM article you can see a comparison between the image allegedly downloaded from the ASU THEMIS image of the day page by Keith Laney and the image currently avaliable on the same page.

Taking a brief moment to examine this image, we first notice that the "NASA/JPL/ASU" text in the top-left seems to be a little blurry, and indeed it is.

As you can see from the comparison image, TEM's image has been blurred while the NASA image has not been.
The next thing that stands out are the rectilinear features present in TEM's image. Let's examine a crop of the two images side by side so that we can note the differences:

The differences however are not startling above, but after equalizing the images (note that these aren't from TEM's JPEG, but from a higher quality TIFF file which they released) we see the following:

And therein lies the city...
Looking more closely at a strip from each image we can see that fine detail has been completely destroyed in TEM's image -- the craters are unrecognizable. The reason for this is excessive blur.


With that in mind, I took the cropped image I'd created earlier (for this article), applied quite a strong blur twice, applied a strong sharpen, then blurred the image slightly to remove the obvious artifacts. What I was left with (equalized) can be seen below. Pay special attention to the arrangement of the blocks (as in the green circle for instance).

That is pretty much case closed on this issue, but there are those for whom this won't be enough, so, let's press on with TEM's article text. The first thing worth mentioning is the following quote:

"We ask that you not download this image unless you plan to run high-level analysis on it and produce a scientific evaluation. Over the holiday weekend, we will produce a step by step tutorial for how to bring out the best results from this data. However, you should know that unless you can obtain software that allows you to use a "decorrelation stretch" technique, you might not get very satisfactory results."

I use DaVinci to process IR images, and I do decorrelation stretches (also note that TEM's step-by-step tutorial was never forthcoming). ENVI is not the only piece of software for manipulating IR images so using DaVinci is not an issue. They go on to state:

"We encourage all interested professionals, especially anyone in NASA with experience working with IR images, and qualified "image enhancement savvy" laymen to use ENVI to duplicate our results. There is a possibility that ENVI may be made available to a wider range of researchers just for this project." -- followed by a link to a TIFF file.

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

On September 5 another article was released by TEM (complemented by another appearance on Art Bell's program to publicize the "city" findings). The first holes in their version of events begin to emerge here. Foremost, we have the following:

"Hoping to see a full color, multi-spectral image (per Dr. Saunders' -- Odyssey Project Scientist -- "preview Web announcement," made just prior to the July 24th release), Laney was so displeased with what was eventually published on the THEMIS web site (a set of grayscale, side-by-side, multiple-band image strips), that he didn't even bother to download -- flatly declaring that "it sucked," when asked about it in the Enterprise Conference and on other message boards across the Net.
Because of this highly visible position, Keith (and a few others who had also publicly criticized the quality of the new image) than began to get a series of interesting "responses" in the Enterprise Conference and in private chats. These "responses" came from two new visitors to our site: "BAMF," and one "Dan Smythe" The former had begun posting about a month before the image release; the latter appeared simultaneous with its release. Both of these new found "friends" seemed to know quite a bit about "infrared image processing," and pretty much goaded Keith into finally going back and processing the full-size TIFF file -- which (according to his own computer log) he had downloaded on the 25th of July, at 10:27 PM EDT."

Interestingly, I was present in that "polite exchange" and nobody goaded Keith into doing anything. However, in the quote above, we see that he "didn't even bother to download the composite THEMIS image the first day (July the 24th, 2002)" and that he instead downloaded the TIFF the next day, "25th of July, at 10:27 PM EDT.". It gets even more interesting because Keith posted an RGB created from 3 bands of the image on July 30. You can see this image on his old site. Below you can see a copy of this image:

You can also view a side-by-side comparison with a 3-band RGB which I created from the TEM image. Both have obviously been converted to greyscale. Notice anything different about them?

Keith's version was quite clearly not created from the TEM image, but rather from the image available on the ASU THEMIS website (the one which Keith claimed not to have downloaded!). But let's press on.

As requested in the first article, I have run a decorrelation stretch on various sets of bands of the TEM image. Here's a comparison of two of them. First, bands 1, 2 and 3, and then, bands 7,8 and 9.


The first thing that really jumps out at me here is the color-bleed on the right-hand side of the images. That tells me that somebody has been playing silly games with the contrast of the strips (the higher contrast having been mistaken for higher detail by some and even a higher resolution). In my opinion, it is contrast manipulation which has caused this effect. In part two, I will go into this in more detail, providing examples and information on how to replicate the effects at home.

The next thing I noticed is that the "buildings" are not in the same places, nor are they the same shapes. Clearly these are just processing artifacts. Take a look at this crop to see that the buildings actually move when seen in the different bands:

Also noteworthy is the fact that the sides of the "buildings" are visible, although the camera points straight down (we should only see the roof of each building). Another puzzler (thanks go to "Diane's 15 Thesis, TEM BBS" for bringing this to my attention) is that the "buildings" are lit from the South-East, as if the sun is shining on them, but this "city" is buried under ice and the image is a representation of heat. This is yet more evidence that the "buildings" are merely processing artifacts.
At this point, I'd like to talk about how the IR hardware on Odyssey works. Essentially, it takes readings of the surface, it cannot penetrate ice, or dust, or anything else to read what is underneath. I posted a brief and very over-simplified message describing this on the anomalies.net message board -- it reads as follows:

"I think you're thinking about it backwards, as has been pointed out before, it doesn't see 'through' anything, granted it can be used to find things underground, but only if they affect the temperature of the surface, imagine a fine layer of dirt with a penny under it, the sun beats down on the dirt all day long and heats the dirt which in turn heats the penny. As the ground is warming up and the penny is still cool, the penny is 'sucking up' the heat from around it, warming up and in turn cooling the dirt, as the dirt cools later, the penny loses it's heat slower and warms the dirt around it. hence, almost all day long, we can see where the penny is. We don't see through the ground at all though, all we see is the effect the penny has on the surface, a fact missed by a lot of people somehow (including Hoagland on Art Bells show)."

With that in mind, a city under the ice is an impossibility. If the buildings were warm enough to affect surface temperature, the ice would melt. If they were warmer than the surroundings (but still below freezing point), the heat would diffuse and the surface reading would be almost constant. The "city", as claimed, is an impossibility.
Back to the article however, and TEM show us nifty little 3D concept renders like this and this. The fact that a camera that points straight down is seeing buildings from a variety of different angles in the same image is apparently lost on them.
They then continue with this gem:

"it is also the only explanation for these buried features we are seeing on the IR composites. It is also the only reasonable explanation for why we're seeing them at all. Long wave infrared imagers have a limited ability to penetrate underground, by detecting surface "hot spots" conducting heat upward from below."

which is essentially correct, but is followed up with:

"But they cannot possibly show the kind of detail we are seeing ... unless ... we are sensing our "city" through a layer of "fluffy, micron-sized dust" essentially transparent to the THEMIS wavebands ... covering another, thicker layer of protective (and also IR transparent) ice."

If the layer was about 10 microns deep, I could believe this, however the "layer" is deep enough to submerge an entire city. This raises a few questions, questions like "where did all this dust come from?", "why doesn't it blow away when the global dust storms occur?", "why can we see boulders on the surface in the MOC images?", "why are there craters in the dust?". The last point is probably going to be the hardest for them to explain away. If this dust really was that "fluffy", any inbound rocks would cut through it like butter and hit the "city" instead. But no, if it is dust, it's definitely not fluffy, and there's no way the hardware on Odyssey is going to see through it. The quote also raises another question, does the camera really see through ice? The answer would of course be no, or NASA would have had a hard time using it to discern the differences between CO2 ice and H20 ice at the poles recently.
Another comparison between processed versions. This time it is claimed:

"When he uses precisely the same processes on the "official" website version the result is absolute, unadulterated, "noise-filled crap"."

Granted the image is not going to be as clean as an image that has been blurred (as was done to the TEM image, which is highly destructive to IR data), but they were indeed left with a terrible image, so terrible in fact that I can't help but feel they deliberately processed it inappropriately in an effort to make their image look better. Here is a comparison between my first attempt at processing the Cydonia image (left) and TEM's image from the article (right).


I'll admit that the TEM version is indeed "crap", but not as much so as the image they are claiming is "real".
Is it a city? No.
Was the image received like this? Again, no.
I have a far simpler explanation.


Stay tuned for part 2, where I'll discuss more of the flaws, inaccuracies and outright misinformation surrounding this image. I will also show you how to create an almost identical fake from the original THEMIS image.
Note: I'm not trying to imply that this image was created by TEM intentionally, but, in my opinion, the city was not present in the image they originally downloaded and rather was a by-product of shoddy noise-removal techniques.

http://no-hype.com/modules.php?op=modload&name=News&file=article&sid=
47&mode=thread&order=0&thold=0