Planetary Imaging with a Webcam and an SCT

Posted by Rod Mollise   05/18/2004 07:00AM

Planetary Imaging with a Webcam and an SCT
[ARTICLEIMGL="1"]The planets are tough. Tough to observe visually and especially tough to image. Those of you who’ve got a few Mars oppositions under your belts know what I’m talking about. I don’t think I’ve ever let a Mars “season” go by without attempting to take a few pictures. Sadly, my usual results have not exactly been something I’ve wanted to show-off to my friends. More likely my photos have been buried in the back of a drawer somewhere safely out of sight. Yes, that bad. Jupiter? Maybe a little easier. But not much. In the past I’ve been satisfied if I’ve wound up with a 35mm slide that shows some vague indications of the North and South Equatorial Belts. The Great Red Spot? Forget it! Saturn? Well, at least you could make out the rings.

I started my planetary picture-taking career in the mid 1960s with a cast-off 35mm camera and an eyepiece projection adapter. Now that’s torture. Mars, for example, is always relatively small, no matter how good the opposition circumstances, and that means enlarging its disk on film by projecting it with an eyepiece. Unfortunately, even with fast film, that makes Mars terribly dim and exposures punishingly long. The high magnification also means the slightest vibration of the scope is highly exaggerated, and that just tripping the shutter is enough ruin a shot. I kept at it, though, trying everything, and occasionally would get an image “good” enough to show some evidence of a polar cap and some vague, vague hints of dark albedo markings. I didn’t feel too bad, though. Heck, wonder-workers like Don Parker didn’t do much better. Even professional film astrophotos of Mars were nearly as blurry as mine.

With the 1990s, electronic imaging became an option for many amateurs. By the middle of the decade quite a few of us were experimenting with video cameras for planetary photography. I, like other amateurs working with video, was easily able to better my film results by using camcorders and black and white surveillance cameras. If you held your mouth the right way, you could even make out Jupe’s Red Spot on my videotapes and frame-grabbed stills. But the results were still not quite there. The problems with video were twofold. First, most vidcams, both camcorders and closed circuit surveillance cams, didn’t offer much in the way of exposure control. Some camcorders available in the early – mid 1990s did have limited shutter adjustments, but if you were using the more sensitive surveillance cams (and sensitivity is important for little Mars and dim Saturn), you usually had to accept whatever came down the video cable, which was often over or underexposure.

The other difficulty was stills. Having a video of Mars to show your friends on TV was fun, but, like other astrophotographers, most of us “astrovideographers” really wanted nice still images to pass around or post on the Internet. Frame grabbers and Snappy image capture cards provided OK results, but some quality was obviously lost in the required analog to digital conversion. Trying to make a noise-reducing image stack by combining dozens or hundreds of video stills using these capture cards (some of which might be able to grab a frame every 5 or 10 seconds) and non-astronomy programs like Photoshop was challenging to say the least.

The big news for planetary workers as the 90s wound down was, of course, the CCD revolution. The amazing results Don Parker was getting with a relatively inexpensive camera showed that the integrating CCD camera was not just for the deep sky stuff. I myself was enthusiastic about obtaining an imager that promised better results than either film or video, and, as soon as CCD prices came down to the 1000 dollar level, I bought a Starlight Xpress MX516 camera. This performed relatively well, producing monochrome images that were much better than what I had been able to do previously with film or video. Jupiter, in particular, began to give up some of the belt detail I’d dreamed about but had been unable to capture with my vidcams.

I still wasn’t overly happy. My planetary images were better, but after seriously using video for five years there were some things I missed. Mainly, live images. I had grown accustomed to focusing using a constantly updating video image on my monitor. The CCD cam, in contrast, given its parallel interface, was able to download a frame to my computer screen once every 5 seconds at best. I could improve on that by going to “focus mode,” but that reduced the frame size to a small spot that would not accommodate the whole disk of Jupiter or Saturn, making it hard to get the planet framed, and, once framed, focused with my moving-mirror-focusing SCTs. Very hard. I also missed the color that I could obtain easily with video. I didn’t have much interest in trying to learn the arcane art of tricolor CCD imaging, either, being more focused on obtaining as many images as possible during times of good seeing than in assembling tricolor “sets.” Since I’m not overly interested in deep sky imaging, these drawbacks of CCD cams for planetary picture-taking meant I wasn’t using my camera much. The last year I owned the MX516 I think I used it twice.

There things would probably have remained if an Internet buddy hadn’t convinced me to try webcams. Yes, webcams, the video devices originally designed for Internet video conferencing. My friend explained that these inexpensive little things were a boon for planetary imagers. Actually better in many ways than the most expensive CCD cams. I was skeptical. How could a hundred dollar webcam best a thousand dollar CCD camera? My friend patiently explained that the characteristics of a webcam that appear to make it inferior to a “real” CCD cam actually mean that it’s better for planetary use.

Webcams do have small CCD chips in order to keep their prices down (some current webcams use CMOS chips—steer clear of these if possible, as CCDs are more sensitive). But this has the effect of providing the planetary imager with a pleasingly large image without having to resort to the dreaded eyepiece projection. In addition, the small pixels of these small chips mean than you get stunningly high resolution, too.

No, off-the-shelf webcams (unmodified webcams) can’t do exposures longer than a second or two at most, but they are sensitive enough that the planetary worker usually doesn’t need longer exposures. And these short exposures come out in a veritable flood of image frames. Five or more every second. Capturing many frames of objects, maybe 1000 or more per sequence, turns out to be the key to getting beyond those blurry washed-out planetary pictures amateurs and professionals have been making since film came to astronomy, you see.

Those of us experimenting with video determined some years ago that the way to capture planets in high resolution is by collecting and stacking frames obtained during moments of best seeing. Being able to build an image out of these best frames means that the imager has the literal equivalent of adaptive optics, something planetary enthusiasts have dreamed about for a long time. This choosing and stacking is made particularly easy by webcams, which deliver that fire-hose flood of images in digital form ready for processing on your computer, no frame-grabbing middleman required.

All this sounded good, but maybe a little too good to be true. I wasn’t ready to invest a dime in a webcam to find out—I just didn’t see how, rational explanations aside, a webcam could improve on my “real” CCD camera. I’d just stick with my CCD setup. But my pal insisted, offering to let me have a Quickcam VC he wasn’t using (these 320x240 format cameras can often be found on Ebay for 10 dollars or less). OK, OK, I’d try it. I started simple, using the software that came with the Quickcam and processing single frames. I was amazed! The little sucker produced images easily as good as those from my best analog videos with little effort. I was able to do even better after I downloaded a shareware program for camera control, K3CCD Tools, and started stacking frames with a freeware offering, Registax.

I was finally convinced that webcams are superior for imaging planets, but was also convinced I didn’t have the right webcam. The main drawback to the VC was its small image format. I wanted a camera that could produce 640x480 frames. Also, I must admit that my VC appeared to be on its last legs. In the course of opening it up to remove the built-in lens (webcams are always used at prime focus or with barlow projection, so you won’t need the lens), I abused the poor thing, breaking its plastic housing. It was held together (most of the time) by a thick wrap of electrical tape. The VC was inserted in the telescope focuser by means of a 1.25” nosepiece I made for it out of a 35mm film can with the bottom cut off. Unfortunately, the hot-glue that held the film can to the camera body kept giving way, dropping the camera to the ground at inopportune moments. It was definitely much the worse for the wear after a couple of months.

Which webcam to choose, though? The most widely recommended model was and still is the Phillips Toucam Pro. Unfortunately, when I got started in early 2003, this camera was very difficult to obtain in the U.S. These days, you can buy Toucams and professionally made 1.25” nosepieces for them from many U.S. astronomy dealers, but at the time I was ready to buy the only option if you wanted a Toucam was to search Ebay daily.

What to do? I remembered SAC, a little company in Florida beginning to make its mark selling professionally modified webcams. A visit to their website revealed that a pair of color models was available, the SAC 4 and SAC 7. The SAC 7, based on the Quickcam Pro 4000, seemed to fit my needs best. It is supplied with a good-looking housing and a nice 1.25” adapter that can be unscrewed to reveal T threads. In addition, it has been modified using the techniques developed by astro webcam guru Steve Chambers to provide long exposure capability. In this mode, the camera can expose for as long as you want, just like a normal CCD camera. I decided this would be handy for Solar System objects like planetary satellites, comets and asteroids.

A final option was the choice between the air-cooled SAC 7 (cooled via a small fan) and the Peltier cooled SAC 7B. The “B” was a little more than 100 dollars more expensive than the standard camera ($499.00 vice $389.00). If I went with the Peltier-cooled cam, I’d also have to purchase an external 12vdc supply to power the thermoelectric cooler. In the end, I chose the B. I concluded that it was just a slightly more versatile camera. I knew that, hot as it is down here in Mobile, Alabama, the Peltier would come in handy if I wanted to take the occasional deep sky shot, and that it might reduce noise for planetary imaging on really warm nights. How to finance a SAC 7B? Simple. I sold my Starlight Xpress MX516 on Astromart. I hated to part with it, but was realistic. The MX516 is a well-made, effective camera, but I wasn’t using it. I had probably had it out about a dozen times—at most—in the two years I owned it.

I placed my order with SAC in late April 2003 and hoped that my camera would be delivered before Mars got much closer to opposition. It arrived in a couple of weeks and I immediately gave it a quick check-out on the Moon. It performed just as advertised-- remarkably well, that is--and I now consider it to have been a steal at its just under 500 bucks price, half what you’ll pay for a basic integrating CCD cam from Starlight or SBIG.

If you’re sure you won’t want cooling and long exposure, a Toucam Pro and 1.25” adapter is a sensible combo at around $150.00 (these cameras are now available from Adirondack Video Astronomy http://www.astrovid.com and Scopetronics http://scopetronix.com). I see that telescope makers are now jumping on the webcam bandwagon as well, with Meade selling a webcam-based “Lunar and Planetary Imager,” the LPI. This camera is based on a CMOS chip, but my early experiments with it show that it seems capable of holding its own with the CCD equipped Quickcams and Toucams.

The camera, of course, is only half the webcam equation. You’ll also need software for image acquisition and processing. The SAC cameras come with a complete package, including the COAA developed program Astrovideo for camera control and FitsX for image processing. Both of these applications are full-featured and well done, but I found I was happier sticking with the freeware/shareware software I’d been using with my Quickcam, K3CCD Tools and Registax. Once the SAC’s drivers were loaded, K3CCD was able to operate the camera just as well as Astrovideo could. Two other items I found useful were an IR blocking filter and a flip mirror.

Webcam chips, like all CCD chips, are quite sensitive to Infrared emission. This is not a problem for black and white imagers using reflecting telescopes. But if you’re shooting color or shooting through a refractor or an SCT, this IR sensitivity is a real headache. The problem for color imagers is “IR bleed.” Infrared winds up being interpreted as the color magenta by your camera and the Moon, Jupiter, Mars or any other bright object will assume a strongly violet hue. This can be removed in image processing, but this is an extra step, and it may be hard to decide which color is “right.” With an IR blocking filter in place, this magenta cast is much less noticeable. You may still have to play with the camera’s “white balance” control to get color where you think it should be, but this is much easier with a filter than without. Off-the-shelf webcams come equipped with built-in IR filters, but most users remove them and discard them immediately. Why? Because they are typically gel-type filters of low quality. Removing a built-in filter also lets you to shoot with no IR filter when so desired for objects—DSOs, for example—that don’t really need IR blocking. Shooting without an IR blocker improves sensitivity for dim objects.

If you use a refractor or other scope with a lens element(s), the Infrared problem becomes more serious than just an annoying color cast. IR is “focused” at a different position from the colors of the visible spectrum. It’s redder than red. This means that even a visually well-corrected refractor (including an APO) may have image softness problems due to the “excess color,” the chromatic aberration, of infrared. This is less of a problem with an SCT or MCT than for a refractor, but I think it is still good practice to use an IR filter with catadioptric scopes. Luckily, Baader 1.25” IR blocking filters are available for a modest price—about $40.00 (try Alpine Astro at http://www.alpineastro.com).

Another accessory I consider vital for the webcam user is a flip mirror. Given the small sizes of our webcam CCD chips and the high focal ratios we tend to use for high resolution planetary imaging—I often shoot at f/30—it can be VERY hard to get a target onscreen. Even a goto scope may not help much unless it can place objects dead-center in the field every single time. A flip mirror makes planet-finding easy. It’s a clever device that’s a little like an off-axis guider. It attaches to the scope (screws onto the back of an SCT or fits into the focuser of other designs), and the camera is either screwed onto the flip-mirror via T threads or inserted into a 1.25” adapter. The flip mirror works very simply. A knob allows you to “flip” a mirror up or down. In one position it sends light to your camera. In the other it sends it up a focus tube to an eyepiece. Flip the mirror up, find the planet in an eyepiece (use a crosshair reticle eyepiece to make the process really easy), flip the mirror down and the planet is on your computer screen. Most flip mirrors have lockable adjustments for the mirror to allow you to fine tune it so that whatever is centered in the eyepiece is exactly in the center of the chip when you flip the mirror down.

The eyepiece focus tubes on flip mirrors are usually lockable, which means you can focus an image sharply on your camera, adjust the focus of your eyepiece until it’s sharp there too, and lock down the focuser so the next time out you can focus the scope for best sharpness in the eyepiece and be assured that Mars (or whatever) will be close to perfect on the computer screen. Flip mirrors are available from a variety of sources, but I’ve found the 1.25” Meade model to be effective for a very modest price. Believe you me, a flip mirror can save time and preserve your sanity.

Once you’ve got your gear assembled, you can start taking pictures and processing your images. How exactly do you do that? Well, come with me on a typical planetary imaging run.

While I could use a larger SCT, I find my C8, a 1995 Ultima 8, very effective, and maybe a little less prone to the effects of poor seeing than my larger instrument. I try to get the U8 out in the backyard and acclimatizing to outdoor temperatures at least an hour before I plan to start my imaging run. Sufficient “cooldown” is critical for good planetary results, as is precise collimation, especially for SCTs. In addition to the scope, I’ll set-up a camping table to hold the computer.

Which is a good time to mention computers. As you may have guessed, a webcam, unlike a video cam, must be used with a computer. A laptop is nice, but not required. Since there’s no reason to leave home for a dark site for planetary imaging, the backyard is fine. That being the case, you can use a desktop instead of a more expensive laptop, a trick I learned from Richard Berry. Using a desktop has other advantages in addition to price. You can use a large, bright monitor, which really helps during focusing, and most desktops will outdo laptops for processing power and storage capability dollar for dollar. Veteran imagers recommend putting a desktop on a microwave or other rollable cart to save the effort of lugging CPU and monitor out every night.

With scope acclimated and computer ready to go, it’s time to take pictures. Mars is fat and riding high tonight, so it’s our target of opportunity. If I haven’t done so already, I mount my flip mirror onto the back of the SCT (I occasionally image with other telescope types, but find an SCT far easier to use than other designs), insert a 2x or 3x barlow into its 1.25” port, and insert my camera into the barlow. I usually shoot Mars at f/30 (f/10 SCT x 3x barlow = f/30) for maximum detail if seeing permits. When the camera’s secure, I plug it into the computer. All modern webcams are USB devices. USB provides adequate speed for image transfer and allows you to conveniently plug-in or unplug the camera with the PC powered up without fear of damaging anything.

My next action is to start my camera control program, K3CCD Tools. When it’s up and running, I select the proper driver for the camera I’m using from a menu (you can use any number of different webcams with K3CCD, since it uses the drivers supplied by the camera maker). I then choose a frame rate for my webcam video, usually 5 or 10 frames per second for best quality, and hit “preview.” Using the flip mirror, I adjust the scope’s aim until the planet is centered on the monitor and, taking my time, focus until I’m satisfied that Mars is as sharp as I can make it. I then recenter the planet if necessary before beginning the exposure. At the high magnification delivered by the combination of a high focal ratio and a small CCD chip, any polar alignment problems or periodic error will be grossly magnified, and you may find that you have to “guide” using the scope’s handpaddle to keep the image centered during the exposure. If you don’t want to fiddle with exacting polar alignment, one of today’s alt-az tracking SCTs can make life much simpler. If your scope features Periodic Error Correction, PEC, you may want to engage it to reduce a planet’s otherwise inevitable drift back and forth across the frame.

Almost ready. Before hitting the shutter button on K3CCD, I bring up the “video source” menu, and adjust exposure and color balance. How slow or fast an exposure? What you should aim for—using K3CCD tools, anyway—is a brightness value (displayed onscreen) of about 150. This provides a good exposure without burning out details. In addition to shutter speed, webcam drivers usually offer gain adjustment, brightness adjustment and gamma adjustment. I’ll advance the brightness toward the top of its scale, and decrease shutter speed to keep the image in the 150-200 range. Gain, conversely, stays close to the bottom to reduce electronic noise in the images. Gamma should also stay near minimum for best results.

If you’re using a program other than K3CCD, what should you aim for exposure wise? Use the shortest shutter speed you can without making the planet intolerably dim. The target should look good on the monitor screen, with no burned-out highlights. It should, in my experience, also look just a little dimmer on the monitor than what you’d consider “well-exposed,” normally. It is far easier to deal with a slightly underexposed image than an overexposed one. Again, always keep the gain as low as possible to lessen noise.

It’s almost time for picture taking! After adjusting the exposure, I make any final focus adjustments I deem necessary, center Mars again, and hit the “expose sequence” button. You can set K3CCD to automatically expose for a given length of time. I generally shoot sequences of about 60-90 seconds at 10 frames per second. This yields .avi files (computer movie files, that is) of about 500-900 frames, giving me plenty of still images to play with. How long you can shoot depends largely on your computer’s hard drive space. Since a single 60 second exposure can consume 500 megabytes or more (that’s right, MEGAbytes), you’ll quickly eat up disk space. It’s usually a better idea to shoot a number of 90 second .avi files space out over the course of the run than to shoot one long sequence. You’ve got a better chance of hitting the really good seeing that way. The drive on my somewhat antique PC will accommodate approximately 15 exposures of this length before I run out of disk space. If you’re imaging Jupiter, you’ll also be limited by the planet’s rotation. Exposing for longer than 90 seconds means that some of the microfine detail on the disk actually begins to blur due to the planet’s fast rotation.

With the .avi files in the can, it’s time to either carry all the gear back inside or to do a little visual observing. I always like to view planets the old fashioned way for at least a few minutes before calling it a night. I’ve been observing the planets visually for nearly 40 years now, and am not about to stop no matter how good computerized cameras get. I will admit I can already see far more with my webcams than I’ve ever been able to see visually.

What’s next? Processing. With the .avi sequences safely on my hard drive, I’ve only just begun the imaging process. All I have is unprocessed video of Mars, and, at this point, a planet will look disappointingly dim and blurry in individual frames. The single frames will, in fact, look so unimpressive you’ll wonder whether you’re wasting your time. You’re not. This is where Registax comes in. Registax, now in version 2, is a freeware program authored by Cor Berrevoets that allows you to do two things: stack the best frames from your .avi files to reduce noise and improve contrast, and process them using its innovative “wavelet” filters to bring out fine detail.

Stacking is very simple with Registax. All you have to do is choose an .avi file on the hard drive, select an area of the image to “track” (because of this ability to track and align images, it doesn't matter if the planet drifts around in the frame), and hit the “align and stack” button. With a planet, I select the entire disk for tracking. When imaging the Moon, I choose a high-contrast crater or other small feature.

Your next action depends on your needs and skills. As above, I just press the “align and stack button.” In this mode, Registax automatically aligns, optimizes and stacks your images, presenting you with a finished picture ready for processing in around 5 – 10 minutes (shorter if you’re using a fast PC, longer if you’re stacking a large number of frames). For sophisticated users, Registax allows adjustment of a variety of parameters and settings, but I have found the “auto” mode more than sufficient to turn out excellent Mars pictures.

Once your frames are aligned and stacked into one picture, the resulting image is displayed onscreen. It will look pretty good, much better than your original frames, and I’d have been more than proud to obtain an image of comparable quality in the bad old days of planetary imaging. But until you run your picture through Registax’s wavelet filters, you’re only seeing a tiny portion of its detail. What are “wavelet filters”? I suspect you’d need a lot more knowledge of advanced mathematics than I have to adequately explain how they work, but, put simply, they are like a series of unsharp masking controls. When you arrive at the processing screen, Registax presents you with a set of six sliders. Each of these controls applies filtering to one “layer” of your image. In practice, moving slider one off zero sharpens small details, slider two works on slightly larger features, and so on. I’m still learning to use this sophisticated tool, but I can tell you that it has done more for my images than any other image enhancement technique I have used. It is positively amazing to move the wavelet controls off zero and watch the semi-fuzzy disk of Mars suddenly explode into a welter of detail!

Following wavelet processing with Registax, I’ll usually save my shot as a .bmp file and transfer the image to a program like Paint Shop Pro or Photoshop for some final tweaking. Actually, Registax 2, offers enough in the way of image enhancement controls—contrast, brightness, gamma, hue, saturation, etc.—that there’s much less need for post-Registax fine-tuning than there used to be. When I’m all done, I archive the best .avi sequences onto CDs. This is very important, since imaging tools like Registax are improving all the time. Who knows how much more detail you’ll be able to pull out of this opposition’s videos with next year’s software?

Yes, I’m turning out planetary shots with my webcams that are far, far better than anything I ever produced with any other media. But that doesn’t mean a webcam is the only path to excellent planetary results. Some remarkable pictures are being shot with digital cameras and CCD cams too. I’m particularly impressed by what Houston’s Ed Grafton does with his SCT and a hardly state-of-the-art SBIG ST5. That said, I think that for most amateurs the webcam is easily the most user friendly and effective tool for obtaining attractive and scientifically useful planetary imagery. Because of the humble webcams, I’ve become so excited about the Solar System again that I’ve renewed my ALPO membership, which I allowed to lapse all too many years ago.

Like any kind of astrophotography, webcam imaging is a learning experience with a steep curve, but there’s plenty of help out there. One good place to look for answers is the Internet Group “QCUIAG” (Quickcam and Unconventional Imaging Astronomy Group, http://www.qcuiag.co.uk/). The members of this group are the true wizards of the webcam, but I’d be happy to help you to the extent I can. If you’ve got further questions about getting started with planetary webcams, or would just like to chart about this new take on the astrophotography game, please feel free to email me (click on my name at the top of this article)