How We Searched for Mars-3

Honestly, it was almost by accident that this news arrived at the beginning of April and Cosmonautics Day. Today I'll tell you how a story that began and ended over 40 years ago suddenly got a continuation in our time. How a simple VKontakte username, interested in Mars, reached NASA. How international scientific solidarity is no empty phrase. And how space is closer than it seems.

We searched for Mars-3.

And we found it! Right on Mars, at the bottom of a giant Ptolemaeus crater, among lifeless wastelands and boulders.

About how we did it — today's story.

It all started on Habr...

No, of course it started on May 28, 1971, when the Soviet research station Mars-3 launched toward Mars. But for me personally, it all started on Habr at the end of November 2012, with one semi-joking comment about NASA: "They found an old Soviet Mars rover there..." At that moment I was confident that all "old Mars rovers" had long been found, since their landing sites were known. But to verify my knowledge, I went to the HiRise website and was surprised to learn that nothing had been found!

That is, the locations of all successful spacecraft were known and all had been photographed by the HiRise camera on NASA's Mars Reconnaissance Orbiter with 30 cm per pixel resolution. But with failures, the situation was less clear. Americans had crashed Mars Polar Lander on Mars, Europeans — Beagle 2, USSR — Mars-6 and Mars-2, and nothing was found, though they searched. But what about Mars-3? It turned out that no one had seen it either.

Our spacecraft, which accomplished a phenomenal, the first in history successful landing on Mars over 40 years ago! It proved this was possible, and did it in practically the same sequence that Americans used to land Curiosity in 2012. Anyway, see for yourself:

So, Mars-3 landed successfully — this is known for certain, since it began radio transmission, which is impossible without deployed antennas. But after 14 seconds the transmission stopped, leaving designers with a puzzle about the cause of shutdown, and space enthusiasts and Photoshop fans — the first photograph of Mars in history.

They did everything with it, applied every filter imaginable, just to indulge imagination and discern the horizon line, mountain slopes and Martian clouds or dust clouds from the raging storm.

But the trick was that this was part of a photo panorama, and to view it in the angle the camera captured it, this frame needed to be rotated 90 degrees. According to designers' plans, the photograph should have looked like this:

(This is a frame from Luna-9, which had much in common with Mars-3)

That is, if you overlay one on the other, it would look like this:

In general, no valuable information was transmitted from the surface, the first Mars rover never set foot on Mars since it was attached to the upper part of the module, but a psychological victory was achieved — we landed our spacecraft on another planet first! The cause of shutdown was never determined. Some suspect corona discharge in the antennas and blame the storm, others suspect the accumulator. In general, future settlers will have something to occupy themselves with.

Then came other expeditions, other successes and failures, and the metal "flower" gradually faded into obscurity. Generations and states changed, the internet appeared, Habr and VKontakte. And my interest in Mars-3 appeared.

Checking the HiRise website, I discovered only an image from 2007 titled "Center of Soviet Mars 3 Landing Ellipse." This was a revelation, since I was so confident in NASA and HiRise's omnipotence that I expected to see precisely where our station stood. A quick internet search also yielded nothing. That is, it was obvious that Mars-3 had never been found. I downloaded the full-format image (1.3 GB), opened it, and understood why in 5 years no one had located the station.

Imagine searching on a rectangle 6 by 20 kilometers for a rounded object 1.5 meters wide. I know many are now thinking that a program should have been written to search for the station itself. But I think such a search is impossible until artificial intelligence is created. Yes, a program could highlight interesting boulders of appropriate size.

But there are thousands of such objects there, since there's a crater nearby from which rocks scattered in a fan pattern.

2 billion pixels, among which you need to find an 8 by 8 pixel object.

Here's a fragment at original size:

Looking at the volume of work, I understood that without collective intelligence this couldn't be done. By that time, the VKontakte group "Curiosity — Mars Rover" already had about 4,000 people, and I proposed to subscribers a space-patriotic quest: find Mars-3. For this I cut the large photograph into 20 strips, uploaded them to the network and proposed to everyone willing to participate in the search. More than a dozen enthusiasts responded, they began viewing fragments and uploading the most interesting finds to a separate album. Much curious stuff was found, but no exact match was shown by anyone:

By its location and shape, the object in the lower right corner could be called with some confidence a braking cone or, in NASA terminology — a heat shield. The Americans agreed with this with the caveat "we probably won't find a better candidate anyway." It was discovered by two search participants: Zaero Ya and Aku Neko.

Hardly had we begun searching when more Google-savvy people quickly found a link on the "Space News" forum where user Imhotep posted an image fragment showing what was likely Mars-3's parachute.

But we finished viewing the entire field and didn't find Mars-3 itself. To verify, I placed all interesting finds on a small-scale map to understand if there was any connection between them. Almost everything missed. Only the parachute and braking cone roughly aligned, but I didn't attach significance to that then.

Then I decided to "dance from the stove" and concentrate on the parachute. That is, I took as a hypothesis that this was indeed it. The problem with the search was that I didn't know the radius over which construction elements could scatter. So I took materials from American missions: Viking-1-2, Phoenix Lander, Mars Pathfinder, overlaid them on Mars-3's landing site and traced the maximum radius over which NASA apparatus scattered from their parachute. I moved the most interesting objects to a separate diagram where I arranged them in approximate correspondence with their position on the terrain. The result looked like this:

(Scale reduced)

But I lacked time and effort for the search. I never finished the diagram, and as it later turned out, the radius I marked didn't include landing elements — our apparatus "scattered" more than American ones.

In my searches and Martian interests, I found Canadian scientist Phillip Stooke. He had once searched for (but incorrectly identified) Lunokhod-2 on LRO images. He suggested the most important moment that proved decisive — the direction of the descent module's flight. He said: "The flight went east." This immediately allowed cutting 50% of unnecessary field and concentrating the search in the optimal direction.

I selected a sector covering four times the radius shown by American experience, and continued methodically viewing the frame. This took several evenings and, honestly, I'd already despaired of finding anything, continuing just from determination to finish. When I announced the mass search, I essentially took on the obligation that people's efforts wouldn't be wasted (though I warned that the probability of discovery was extremely small). In any case, I had to verify so I could confidently say "Finding is impossible" or "We found it!"

The result was — "We found it!"

Initially I concentrated my search on the module, which should have been identifiable by its cross-shaped "petals." I recognized the object by these. And the soft landing engine, coupled with the parachute container found nearby, turned out to be a valuable bonus. I based this on film, where this detail "disappears" before landing, so I hadn't searched for it. In fact, the final stage proceeded in this sequence:

Found it! But what next?

I could have written about this on Habr or on the "Space News" forum, but there are already plenty of searchers there who see Mars-3 in every bump and a parachute in every sand dune. Proof was necessary.

Here Phil Stooke helped again. He looked at my find, agreed it was very interesting, and suggested identifying several other objects in the vicinity that resembled the descent module in size and shape. This was needed to show my find was the most convincing.

I did this, and got a long list:

(Clickable. 1 MB)

More precisely, there were two, but I think one is sufficient for example. Number one is Mars-3 itself (the most convincing candidate). Of course, this selection isn't entirely representative since it was compiled when I already knew the correct answer, but I tried to select the most similar ones. Fortunately there weren't large rock piles like in the photo above in the search sector.

But this was just one argument for my find. I needed more convincing proof, but I couldn't do it myself. A repeat image of the terrain at different times was needed to determine object shapes by shadow changes. Information about the cable length connecting the parachute container and soft landing engine was needed. This last point could not only confirm but refute my guess, which aligns well with Popper's criterion, which all scientific theories must satisfy. The length observed in the image corresponds to 4.8 meters. If a discrepancy of even a meter appeared, claims to discoverer status could be forgotten.

Here a question arises: how could a thin cable be visible in images with 25 cm per pixel resolution?

Looking at Opportunity's photograph, you can see the shadow from its mast.

Although in reality its thickness doesn't exceed 15 cm:

That is, HiRise sees what it shouldn't see.

Though even imagining a 15 cm cable in this role is difficult, I assumed sand could accumulate along it, and we're seeing not the cable but a small ridge that formed under wind action. Further research showed even these assumptions weren't necessary.

HiRise only imaged this area once, and information about cable length in scarce open sources doesn't exist. The most informative source — the book "The Difficult Road to Mars" by V.G. Perminov (incidentally published with NASA funds in English). But needed information wasn't there either.

So I contacted Professor Alexander Basilevsky, heading the comparative planetology laboratory at the Vernadsky Institute of Geochemistry and Analytical Chemistry (GEOKHI). He was mentioned on Wikipedia in the Lunokhod-2 article, so I thought he'd be interested in the Martian theme. Initially he was skeptical: "Argument is needed — why you think this object is what you think. The argument one acquaintance used: 'Only the blind don't see this' — doesn't work here."

I prepared a diagram showing everything I "saw" and how:

And convinced him! Of course, being a scientist, Alexander Tikhonovich emphasized his doubt and always indicated "Mars-3 elements as Vitaly saw them" in correspondence. But he wrote to HiRise camera group leader Alfred McEwen.

The American's response was quick and unexpected: "We will re-image this location to see if we can learn more."

Just like that! I don't know whether my illustrations were so convincing or Basilevsky's international authority so high that NASA agreed right away to adjust its orbiting apparatus to photograph the spot where something caught a simple kid's eye from Soviet Russia.

While NASA aimed, Basilevsky "infected" Arnold Selivanov with my find — head of the RKS scientific-technical center and one of Mars-3's creators. He also quickly overcame initial skepticism and made efforts to verify information from his side. The notorious cable length on the soft landing engine had to be established. This could only be done by penetrating NPO Lavochkin's depths. Under Selivanov's direction, one of NPO's engineers — Vladimir Molodtsov — undertook this feat.

For me a lull followed: HiRise wandered near-Martian expanses, Vladimir — NPO Lavochkin's expanses.

Trying to extract whatever information from available photographs, I calculated the module's height based on shadow length and illumination angle. I got 78 cm. Mars-3's actual height — 75 cm (height to width ratio when unfolded 1/2). A small point in favor, but main proof had to wait.

Weeks passed, April approached.

First word came from our side, from Vladimir: "Chains connecting the soft landing RDTT to the parachute container are 4.52 meters long."

Bingo! 4.8 on Mars, 4.5 in drawings. Error within 1.5 pixels. Plus the visible 0.3 m fits perfectly for the engine. This was simply a magnificent gift. We hurried to share with Americans, and received no less positive news: "We have a new image and are working on special noise removal and sharpening of both images."

Days later came the ready result:

(Clickable. 3 MB)

By the way, the image they got is lower quality than the 2007 frame due to higher atmospheric dust, but the color camera field captured the module itself. However, this gave no color information, since if painted any color, the paint would fade under ultraviolet rays and the surface would be covered with a dust layer.

But now we can compare "before-after" and view the field under different lighting angles.

New images look paler also because the sun is higher at zenith, shadows are shorter and everything seems flatter.

Then we spent another week coordinating the press release. This incidentally highlights NASA's ethics — they could easily write independently, even mentioning us. Though perhaps this is just international cooperation principles. Unfortunately, they couldn't place everyone's names who helped me search on NASA's site, but they did this on NASA JPL's site.

Our activity, searches, proof and archive excavations results — on NASA's site.

Now to business: Mars-6 still hasn't been found!

I'd like to offer readers to contribute to Soviet cosmonautics history. One could rush in crowds to HiRise, recheck images and find something suitable. But more effective for popularization would be creating an Android and iOS application allowing viewing separate image fragments and highlighting the most interesting areas. Something like the "Fourth Planet" project. Where else to address such ideas but Habr? I figured one image 25K pixels by 80K would give 8,333 fragments 600x400 pixels. For Mars-6 such frames were made several, so there's plenty. I think if such an application were made, it would help searching for American Mars Polar Lander and European Beagle 2 too.

After all, it's better seeking lost spacecraft than casting swine before pearls! I'll try attracting support from Roscosmos, if anyone there thinks about popularization, or even NASA. Let's open a separate comment thread on this topic and discuss details.

And yes, none of this would exist without my employers, who invited us to Mars and turned a blind eye to work time I spent staring into HiRise's gray field.

And happy Cosmonautics Day to everyone!