With a Camera to the Red Planet
The engineer who drove the Mars rovers recalls what it was like to see the first images arrive, and how he worked the cameras.
When John Wesley Powell and his nine men pushed their four boats out into the roaring Colorado in 1869, they had no idea what lay downriver. They set out with the knowledge that they might not return—and several did not. As I reflect on the last decade's adventure of designing, building, testing, launching, and operating two complex and hardy robotic space vehicles on Mars, I cannot help but wonder if we were just as naive when we started out. Our lives were not on the line, of course, but we certainly ended up risking our health, marriages, family relationships, friendships, to some degree our scientific careers, and sometimes, it seemed, our sanity.
Spirit and Opportunity were not the first missions to land on Mars, rove its surface, or transmit images of its ruddy alien landscape. The two Viking landers (1976-1982; 1976-1980) and the Mars Pathfinder and its Sojourner microrover (1997) beamed back tens of thousands of images that revealed Mars’ surface as rocky, dusty, and yet strangely familiar. The difference between the views taken by the earlier missions and Spirit and Opportunity is that between “acquiring images” and “taking photographs.” The former is a technical, science-driven, and resource-limited activity. Every space mission—whether human or robotic—carries a camera to pass images of alien worlds to those back home. It is not easy to take these pictures or to send them home: Spacecraft and their instruments are complex, sometimes finicky to operate. There is often little time to take pictures. Even scarcer, usually, is the bandwidth necessary to transmit good quality pictures to Earth.
Our Mars rover photography team has had the luxury of being able to devote far more bandwidth and time to picture taking than ever before. In addition, our cameras have far better resolution. As a result, we can be photographers and artists while at the same time gathering required scientific and engineering information.
More than 100,000 pancam images traveled from Mars directly into my laptop
My interest in photography began when my parents bought me a Pentax 35 mm SLR camera when I was a kid. I spent a lot of time shooting the outdoors with my friends in the high school Photography Club, becoming fascinated with the interplay of light and shadow in the environment—and how a photograph could be framed and composed, like a musical piece, to tell a story to the viewer in a certain way. At the library I soaked up Marcel Minnaert's book The Nature of Light and Colour in the Open Air and checked out picture books about 19th- and 20th-century landscape photographers. When I figured out how to attach a camera to my telescope, I was hooked. Space was the ultimate landscape. I knew then that I wanted to get into astronomy and space exploration. Little did I realize then that I would have the opportunity to take some of the most spectacular photographs of the Martian landscape.
The Expedition Conceived
The idea of a robotic geologist roving the rusty sands of Mars was the brainchild of a small group of planetary scientists and engineers lead by Steve Squyres in the late 1980s and early 1990s. When Steve asked me to head a team that would build and operate the rovers’ panoramic color cameras, or Pancams, I personally took responsibility not only for making sure the cameras would shoot great photographs, but that every Pancam picture or panorama that we released would be of the highest quality. Every single one of the more than 100,000 Pancam images has traveled from Mars to Earth and then through my laptop computer (sometimes to the frustration of my colleagues) before being posted or printed.
The package of cameras and other scientific instruments that eventually morphed into the payload carried by Spirit and Opportunity was first proposed to NASA in 1995 as a set of experiments designed for a Mars lander to be launched in 1998. While that mission was scrubbed, NASA gave the green light in 1999. The launch date was implacably set by Newton’s laws of celestial mechanics. About every 26 months, the Earth and Mars align in such a way that a rocket can use the minimum amount of energy necessary to cross the gulf, which requires a smaller and less expensive vehicle. The next opportunity coming up for the rovers was in a “window” extending from June to July 2003. We would work at the Jet Propulsion Laboratory (JPL), a NASA research facility near Los Angeles. The clock was ticking.
By early 2002 we had settled on our near-final payload instrument design. Each rover was outfitted with nine cameras; a tenth was carried by the lander to take pictures of the landing site during descent. Two pairs of the nine rover cameras were very wide-angle, black-and-white stereo instruments--called hazard avoidance cameras, or Hazcams. Mounted low to the ground at the front and rear of the rover, they helped engineers characterize potential obstacles while the rover was in motion. Two more pairs of cameras were mounted atop a 1.5-meter-tall mast at the front of each rover, one of these pairs being a set of wide-angle, black-and-white, stereo navigation cameras or Navcams, which provided stereo coverage of the terrain around the rovers. Another camera set on the rover’s arm was a small microscopic imager, designed to take black-and-white images of an area only about three centimeters across at a scale comparable to a geologist’s hand lens.
The two remaining cameras mounted on the mast were the Pancams, designed to obtain high-resolution images of Mars in colors ranging from ultraviolet--somewhat deeper into the blue spectrum than the human eye can perceive—to infrared—a bit more red than human color perception- Each Pancam had a small, eight-position filter wheel on the front that enabled us to obtain images in different parts of the spectrum, which we then combined into color images. A typical color image is usually the combination of three separate images, shot through red, green, and blue filters respectively. We downlinked all three, and then combined into a color composite by computers back on Earth. Consumer digital cameras today do essentially the same thing except that they have internal filters so that the images are processed and combined into a color composite all with the snap of a shutter.
When photographing another world, it is important to have some earthly reference. While everyone had an idea that Mars’ surface and sky would be reddish, going by telescopic and early orbiter camera views, no one knew exactly what shade or shades would dominate the landscape, nor whether there would be rocks and soils of different colors. Back in the 1970s, the Viking team realized that if they took a picture of an object of known color along with the rest of the scenery, then they could deduce the true colors of the unknown rocks, soils, and sky. When the Viking lander cameras were in design, the scientists and engineers created a sort of calibration plate containing painted grayscale and color patches of known brightness and hue that could be carried along with the camera.
A decade later, the design teams for the Mars pathfinder modified the Viking design by adding a central post in the middle of two calibration plates—similar to the “gnomon” in a sundial—that would cast a shadow across the target. Only scattered light from the sky would fall within the shadow, so images of the sunlit and shadowed area of the target would allow the sky’s color to be measured and removed from the color of the surface rocks and soils.
When we started designing the Pancams for the Spirit and Opportunity rovers, we tweaked a few things from the Pathfinder 1, but kept the basic rings, colors, and shadow-post design to create a simple, functional, no-moving-parts device that would help provide important calibration data for the Pancam. We called it the MarsDial.
Putting the Cameras Through Their Paces
Perhaps the most grueling and intensive activity that the camera team went through before the launch was a series of laboratory measurements during most of 2002 and the first part of 2003 that were carefully designed to test and calibrate the instruments. First by themselves, then after assembly on the rover, each camera was subjected to the shocks of launch and landing and the same kinds of extreme cold and near-vacuum conditions that they would encounter on Mars. To play it safe, we had to prove that the instruments would work in even harsher conditions than the ones we expected them to encounter. A few dozen scientists and engineers from the team, as well as a handful of wide-eyed young Cornell students, spent months working in shifts to plan and execute round-the-clock tests inside cold, loud JPL basement labs, surrounded by vacuum pumps and test chambers. We carefully measured how the cameras performed and focused them on rocks and other targets to ensure we had the necessary accuracy and sensitivity. We took pictures of an empty Starbucks Frappuccino bottle when we needed a green imaging target. We practically lived on corn chips and the wonderful smoky salsa at Dona Maria's, a little Mexican restaurant not far from JPL.
The calibration tests culminated in a series of thermal vacuum measurements: each complete, fully-assembled rover was subjected to Martian temperatures and pressures inside a large vacuum chamber. It was a frantic pace, building and testing more than 40 different cameras in only a few months—20 cameras on the two rovers that would go to Mars, more for the two rovers that would be used for testing on Earth, and all kinds of spares for both.
Sending photographs back from outer space is a complicated task. Typically, spacecraft will acquire their pictures and other data, encode them into rather feeble radio signals, and transmit these to the Deep Space Network (DSN), a series of large radio telescopes spread around the Earth. This direct-to-Earth communication capacity is subject to important constraints. First, the Earth must be visible to the rover's transmitting antenna, often a problem because Mars spins and the Earth is only "up" for about twelve hours out of each twenty-four hour thirty-nine-minute Mars day, known as a sol. For half of each sol, direct-to-Earth communications are not possible. Second, such communication requires a lot of power. Power spent by the rovers transmitting data means less power for driving, taking pictures, or other operations. Like everything else, communication time is a resource that has to be carefully managed.
All during Spirit ’s entry, descent, and landing phase on the evening of January 4, 2004, the Earth was in constant view of the spacecraft; shortly after the landing and before the rover finished unfurling from the lander, however, the Earth set below the horizon, making it impossible to diagnose the rover or to send the first pictures and other data back home by direct communication.
To circumvent a blackout period, we attempted a bold experiment
To prevent such a blackout period from occurring, the team attempted a bold experiment. It was known that if all went well, just after the rover finished unfurling and acquiring its first pictures, the Mars Odyssey spacecraft, a NASA orbiter some 400 km overhead and studying Mars since 2002, would pass right over the Spirit landing site and be visible above the horizon for about ten minutes. The rover had been preprogrammed with the time of Odyssey ’s transit. At that time it would attempt to uplink the images and other data directly to the orbiter rather than to the Earth. Odyssey is equipped with a special receiver cleverly designed in advance for just such an opportunity. If Odyssey received the data, it would relay the images and other information back to Earth. This, essentially the first attempted satellite phone call from another planet, marked an important first step in NASA's plans to use "local assets," such as orbiters, as part of a Mars communications infrastructure.
If the experiment worked, this power to communicate with the rover after the Earth had set could double or triple the amount of data that we could send back compared to the direct-to-Earth method. A greater volume of data could be sent to and stored on Odyssey because the orbiter is much closer to the rover, which means the radio signal is so much "cleaner."
Spirit ’s first color postcard home, acquired the morning after the landing, covered a swath of reddish-brown ground extending from the airbags at the base of the lander out to the horizon. I was elated when the image popped up on my screen because everything appeared to have worked. The sequence of 36 images--a mosaic three images wide by four images tall, in three filters--went through without a hitch. We used MarsDial images to calibrate the data, and the colors came out close to what I was expecting going by what I had seen of Mars through a telescope. The mosaic-making and analysis software built so painstakingly by Cornell team members stitched the images together like a charm.
I zoomed in on pitted rocks and wondered if wind or water had shaped them
At the same time, though, I got this sinking feeling. The image was flat and boring, only a few rocks and nothing of interest in the distance. I loaded it into our Pancam image-viewing program and zoomed in until I was finally at the camera’s full resolution. At that scale, which is difficult to reproduce on the printed page, I could begin to see that certain rocks were angular, others pitted, and some smooth and rounded. Had wind or water shaped them? Faint hills and mesas indeed appear on the horizon, but are partially obscured by dust and distance. Several bright circular depressions, probably impact craters, were clearly distinguishable. Nearby, next to the airbags, which had cushioned the landing, I could see what looked like very strange scratch marks left in the soil--as though a carpenter's plane had been dragged across the ground, lifting and curling pieces of soil like wood shavings--as the bags deflated to retract into the lander.
Stitching It All Together
During the next week, while the engineers prepared the rover to drive off the lander, we scanned the cameras around to build up our first 360° color panorama of the landing site. This was dubbed the "Mission Success" panorama, because part of one definition of "success" for Spirit was to acquire and downlink such a comprehensive opening color image. Acquiring color panoramas is a painstakingly slow process for the rovers, as the field of view of the Pancam is only 16° x 16°, or roughly equivalent to the view in a 35-mm camera with a 100-mm lens on it. Images must slightly overlap, which entails taking 27 separate images to cover one 360° scan around the rover.
Usually we have to stack at least three of these scans on top of one another so that we can cover the scene from just above the horizon down to the rocks and soils close to the rover. That is 81 images, just to obtain a panorama through a single filter--on average, about 90 minutes’ work. But to obtain a panorama in three colors plus stereo vision, we have to quadruple those numbers. A 360° color panorama consists of 324 separate images and requires more than six hours, usually spread over several different sols, to collect. Figure additional time and images if more color filters are needed. A pan can take three to six sols to acquire, and sometimes a week or more to get their 300 to 500 million (or more) bits of data back to Earth. For the initial Spirit panorama, we only had a few hours’ imaging time per sol. The engineering activities to prepare the rover to drive off the lander had priority, so it took more than four sols slowly to accumulate the pieces of the Mission Success pan.
For more than a century panoramic photographers have known that seams or sharp color changes make it very difficult to impart a perceived sense of visual reality to a landscape. For purely scientific-analytic purposes, seams and other image artifacts may be acceptable or even desirable. But to give the scene a sense of reality, we removed the artifacts from many of our large rover color composite panoramas and attempted to simulate the view a human observer would take in if all the images in the mosaic were registered at the same moment. The successful simulation of such a view rests on a combination of science, past experience, and some artistic license, although the colors and contrast are never "faked."
Unlike the initial postcards, Spirit ’s spectacular Mission Success panorama provided a much more complete view of the geological setting of the rover's new home, which proved to be in a gently rolling plain, about five percent of its surface covered by rocks bright and dark, angular and smooth. Some rocks had bright wind-tails that helped identify the direction of the prevailing wind. Many hills and mesas rimmed the horizon.
Our cameras could faintly discern brighter and darker bands on the ground as we looked out over the plains toward the hills--the dark streaks that we had seen in orbital images, now viewed up close. To the north a ridge marked the crest of a large crater, which we could see from orbit had bright material on its floor, but could not tell what it was from the rover. We could see several more of those enigmatic, bright-floored hollows, often ringed by rocks, but showing few rocks inside. And we saw more of those dirty airbag parts at the edge of the lander, what used to be pristine white fabric now covered by fine-grained, reddish dust--a dirty, dusty, alien environment. While these images from Mars evoke a sense of familiarity, it is only an illusion. It is 30 to 50 degrees below zero on average; the air is almost entirely carbon dioxide, with only a trace of oxygen; and it has not rained in something like two to three billion years, if ever. The rocks have been carved and molded for billions of years by wind-borne sand and dust grains. The ground is peppered with circular holes, impact scars formed when asteroids or comets crashed into the planet long ago. Spirit is driving on and over terrain that is as old or older than the oldest places preserved on the surface of the Earth, Some of the rocks we have examined are probably three or four billion years old, and Spirit driving past may have been the most interesting thing to happen in some of these places for the past billion years or more.
Six Wheels in the Dirt
On Spirit ’s twelfth day on Mars, we completed the final 35 cm of our entry onto Mars, rolling off the lander and onto the surface. When the first images came down looking back on the lander from our new vantage point on the ground, applause and tears and hugs and glee burst out all over again. We fired up the Pancams for some color imaging of the lander and shot what has turned out to be one of the most stunning and emotional postcards of the entire mission. This so-called "Empty Nest" panorama shows our lander, now just a dusty, dirty, and empty cocoon, framed against the backdrop of the plains of Gusev and the distant hills on the eastern horizon. Many of the engineers were interested in seeing firsthand the detailed final configuration of the lander and the airbags. They knew that it had all worked, of course, but they also knew that future missions might benefit from a high-resolution, after-the-fact study of a successful design. Many of the scientists enjoyed the detailed color views of the soils and rocks near at “hand” and the plains and hills in the distance. We were still trying to figure out what kind of place our new home really was.
Pot of Gold
Spirit made it to what became known as the West Spur of the Columbia Hills in July 2004, around SOL 157. During our drive across the plains, we took color images of the hills at higher and higher resolution to help us target the safest and most interesting parts to eventually drive up into and explore. Images of the West Spur, a smaller hill connected to the rest of the Columbia Hills complex, were used to choose it as our mountaineering entrance point partly because it appeared to be relatively easy to climb and partly because it is one of the closest places in the hills to Spirit ’s landing site.
The West Spur turned out to be a fascinating choice, because almost literally as soon as we started climbing we came across a bizarre and different kind of rock. We quickly found that indeed the geology and mineralogy were different from the plains over which we had driven for nearly six months. In fact, the first interesting rock we found was so different that we named it Pot of Gold because we surely must have found the end of the rainbow.
Pot of Gold is a small rock, only about 10 cm across, sitting inside a much larger, round hollow of bright dusty soils near the base of the West Spur. Close-up examination with the cameras revealed what looked like a strangely eroded and etched surface, with "stalks" of rock jutting out and weird little nodules on their ends. It was unlike any rock we had seen in Gusev. Some of the patterns looked familiar to the geologists on the team who had experience working with heavily eroded rocks on Earth.
The Mössbauer spectrometer detected the presence of coarse-grained hematite, the iron-oxide mineral discovered from orbital data that led to the choice of Meridiani Planum as the landing site for Spirit 's sister, Opportunity. Here in the Columbia Hills, as at Meridiani, it provided a beacon of sorts, indicating that liquid water may have been involved in the alteration and weathering of these rocks. This was a truly special moment, because we had found evidence, in a single rock, for both physical and chemical alteration that had a good probability of involving water.
It was momentous for the team when Spirit finally reached the summit of Husband Hill after a year of zigzagging and scrambling
As the rover continued to climb the West Spur, we found that the rocks we encountered in the hills were not pristine volcanic basalts, but weathered and altered, containing minerals formed by the action of water and heat. The Columbia Hills really did appear to be remnants of a more distant period in the Martian past, when near-surface water may have been more abundant and the climate correspondingly different.
We took a path up the flanks of the tallest peak, Husband Hill, the rover often scrambling to get up the slopes, dislodging rocks and digging deep wheel tracks, or mini-trenches, as it toiled uphill.
View from the Top
Finally, after nearly an Earth year of taking photographs while zigzagging, slipping, trenching, scrambling, and generally rugged climbing, Spirit reached the summit of Husband Hill, a broad plateau about 100 meters above the plains where we had landed. While Husband Hill would be considered a rather low hill on Earth, for the rover and the team, reaching the summit was momentous. From our vantage point high above the plains we acquired our largest Pancam panorama during Spirit ’s mission, a behemoth 653-image mosaic in five filters covering every square centimeter of terrain from the rover deck out to the horizon: a magnificent view, with many alluring prospects to explore.
Sometimes I go back and look at our early pictures from the lander. I cannot help but feel like I do when I look at baby pictures of my children. Ah, we were so young, so naive. We were incredibly stingy with our resources, often using fewer color filters than we really wanted or compressing the images to levels a bit harsher than we should have. We were paranoid that every picture, spectrum, and chemical measurement could be our last. If not for the incredible ingenuity and sleuthing of so many clever people working hard together to solve a difficult problem, we would have lost the mission right at the beginning. Since then we have dodged a few more bullets and been on the receiving end of some plain old luck. It is almost impossible to believe that our plucky little machine is sitting there on top of those hills on Mars. And we still do not know when our photographic and scientific adventure will end. Stay tuned.
Adapted from Postcard From Mars: The First Photographer on the Red Planet by Jim Bell. Published by arrangement with Dutton, a member of Penguin Group (USA), Inc. Copyright © Jim Bell, 2006.