Meteor observation has a long tradition in amateur astronomy. It is one of the few areas where amateurs can still contribute to scientific advancement in astronomy today. This is primarily due to the nature of meteors, which are virtually unpredictable. While it is fairly easy to estimate how many meteors will be observed on a given night, when and where a meteor will appear in the sky is completely random. Furthermore, observers are needed at different longitudes and latitudes to fully cover meteor activity. Observations from a single location are (at least in the optical range) limited to nighttime hours. Additionally, a meteor shower is only observable when its radiant is above the horizon. Finally, there are only a negligible number of astronomers worldwide who professionally study meteors. They often can only engage in the analysis of observations and modeling of meteor showers, without actually observing themselves.
Among amateur astronomers, there are hundreds of meteor observers in a variety of countries worldwide. They can gather data year-round and around the clock. Nearly 15 years ago, the International Meteor Organization (IMO) was founded. It defined observation standards that enable the combination of observations from different observers within the framework of global meteor stream analyses. Since that time, the IMO has been collecting data from national meteor observer groups and individual observers, facilitating and maintaining contacts with these groups, and promoting various observation methods.
For someone who has not delved deeply into meteor observation, it may be surprising that the fundamental observation method has changed little over the past hundred years. The majority of all data is still collected by visual meteor observers. The night sky is scanned with the naked eye from a dark observation site. Individual meteors are either just counted (counting) or plotted on special gnomonic star charts for more precise measurement later. The success of this observation method is not only due to the fact that no technical equipment or special conditions are required, but also to the properties of the human eye, which is a very effective meteor detector. We can monitor a large field of view with good magnitude limit simultaneously. Other observation methods have (yet) not achieved this significance. Meteor photography, for example, can only detect the brightest meteors. Using a bright lens and sensitive film, one can only record meteors with a brightness of 0 or +1 mag, and even the large Super-Schmidt cameras of the fifties could only capture meteors up to the third magnitude class. That is still three magnitude classes less than what the human eye can achieve with a significantly larger field of view!
In the second half of the 20th century, radar observation of meteors became the first choice for many professional astronomers. An active backscatter radar system that records the ionized meteor trails is expensive, but it allows automated observations round-the-clock. Furthermore, radars can detect much weaker meteor trails than visual observers, registering significantly more meteors. However, it is very complicated to determine essential parameters of a meteor—such as its direction, speed, and brightness—with great precision from radar data. In the amateur version of radar observation, passive forward-scatter observation, data evaluation is even more complicated. For this reason, amateur radio meteor observations are only suitable for monitoring general activity levels and registering outbursts.
Video observation is the newest meteor observation method. Image-intensified video cameras were first used for meteor observation in the sixties. Among amateurs, Japanese and Dutch observers were the first to use video systems from the late eighties. To get to the point, video observation combines many advantages of other observational methods in the optical area. However, the technology is quite expensive and dependent on electricity. A video camera can register meteors just as effectively as a visual observer by monitoring a large field of view with a good magnitude limit. It is also more objective and not prone to observational errors. A video system can determine the position, brightness, and angular speed of a meteor with high precision, without getting tired or falling asleep. Video systems are also interesting for observation sites that cannot provide a perfect night sky, as they are significantly less susceptible to scattered light than a visual observer.
The most important component of a video meteor camera is the image intensifier, usually of the second or third generation with a microchannel plate (MCP). Even today's best CCD cameras without image intensifiers are significantly inferior, as they are less sensitive than intensified cameras and can only capture the brightest meteors. Typically, a bright photo lens is mounted in front of the image intensifier. The lens focal length significantly determines the properties of the meteor camera: With a standard 50mm lens (aperture 1.8), you achieve a field of view of about 30 degrees in diameter with a stellar magnitude limit of 7 to 8 magnitudes. A wide-angle lens (e.g., 1.8/28 mm) delivers a larger field of view (about 50 degrees) with lower magnitude limit (approx. 6 mag), whereas a telephoto lens 1.5/100 mm records stars beyond the 9th magnitude at a field of view of only about 15 degrees. To monitor general meteor stream activity, a large field of view is usually preferred, while for multi-station observations of meteors, telephoto lenses are preferred, providing greater spatial resolution and positional accuracy.
At the back of the image intensifier is a regular CCD video camera, whose recordings are captured on videotapes for more detailed analysis later.
Automatic Meteor Observation with Image-Enhanced Video Cameras
What do you need for automatic video meteor observation?
An image-intensified video meteor camera consists of three parts. The most important is the image intensifier, which should have sufficient gain (5,000), a large photocathode (preferably 25 mm), and as little noise and image distortion as possible. Image intensifiers of the second generation or higher are usually well-suited. Sometimes, used military image intensifiers are offered at reasonable prices. Often, the image intensifier is delivered together with a high-aperture lens; otherwise, one can also use normal 35 mm photo lenses. The video camera for recording the screen doesn't need to be particularly sensitive. More importantly, the entire phosphor screen of the intensifier must be sharply imaged, which often requires a coupling optic with somewhat longer focal length. Recently, new CCD video cameras have entered the astronomy market that are significantly more sensitive than their predecessors. Thanks to higher light efficiency, the highly sensitive CCD chip over a wide wavelength range (e.g., Sony ExView HAD), and the extremely low-noise readout electronics, these cameras can also be used for meteor observation without an image intensifier. Although they are still inferior to an image-intensified system, they are much cheaper, more robust (e.g., insensitive to the moon and twilight), and have a long lifespan. For automatic data analysis with MetRec, a PC with DOS (or Win95/98) is needed. A Pentium PC with 200 MHz and 16 MB RAM is the lower limit, and with 500 MHz and 32 MB RAM, one is already on the safe side. MetRec supports the frame grabber family from the Canadian company Matrox (http://www.matrox.com) and is specially tailored to the Meteor-II frame grabber, which costs just under 700 Euros new. The software can be downloaded from the MetRec homepage www.metrec.org. Amateurs can use the software for free, while professional or commercial users must pay a registration fee initially. Upon request, the source code of MetRec is also made available.
For a long time, data analysis was the bottleneck in video meteor observation. If you have a meteor camera, you can record the night sky for hours and capture many meteors. However, finding and measuring the meteors is a complicated matter. When amateur astronomers began video observation, PCs were still in their infancy. At best, they could support position measurement and coordinate transformation, but finding meteors on the video tapes required observers to watch the tapes. Position measurement also had to be done manually, as frame grabbers were not yet available. Many interesting questions could have been addressed if the evaluation of video data were more effective: monitoring meteor shower activity throughout the year regardless of the moon, temperature, and clarity, studying small and large meteor showers, registering meteor outbursts, and detailed studies of specific meteor showers (brightness distributions, light curves, afterglows, etc.), fireball monitoring, studying telescopic meteor showers consisting only of very faint meteors, and much more.
The first attempts to write software for automatic meteor detection date back to 1993. At that time, the author developed a program capable of searching for meteors in the video data stream in four passes. In the end, efforts were only partially successful due to the limited computing power of PCs and the low data transfer rate of frame grabbers. The first generation of image intensifiers, which were initially used, were also problematic due to their high noise. Three years later, an American amateur wrote a program that could detect meteors in real-time. He connected the computer to his meteor camera and was able to record meteor images live through the night. Although his system was never regularly used, it showed that PC-based data analysis in real-time was no longer impossible. Meanwhile, we also used more and better meteor cameras, increasing the demand for efficient evaluation software. For this reason, the old software package was completely rewritten by the author. By mid-1998, the first version of the new software MetRec was ready.
At first, the goal was simply for MetRec to register meteors in the video signal in real-time and store their flare-up times. It was shown that the software performed this task well. In high-quality recordings, MetRec registers about 80% of the meteors and only misses the weakest ones. This is comparable to the detection performance of an observer inspecting the video tape on a television in one run.
Once this task was solved, the software was further improved and expanded. The focus shifted from a pure meteor detector to a software package that efficiently conducts complete data analysis online. Routines for coordinate transformation and measurement of brightness and angular velocity were written. The automatic identification of meteor showers and the storage of meteor images and short image sequences were added. Since the aim was to have an automatic meteor camera at the end of the development, the main criterion was that everything should be done with as little human interaction as possible, with only a slight loss of accuracy compared to manual measurement.
By early 1999, MetRec was capable of observing meteors fully automatically. Since March of that year, a meteor camera has been operated regularly in Aachen on every clear night. In July 1999, a second station was added near Potsdam, and additional stations provided supplementary observations on selected nights. By mid-2002, the AKM video camera network had grown to eight stations worldwide that regularly monitor the night sky, and another five stations that contribute data irregularly. MetRec is also used by various astronomical institutes. The software greatly aided in analyzing the Leonids in 1999 and 2001, where detailed results of the meteor showers, such as activity profiles and radiant plots, became available shortly after observation.
A typical automatic meteor observation takes about 15 minutes of manual work and yields between about twenty (spring) and one hundred (fall) meteors on a clear night away from major meteor streams. The camera is set up in as dark a location as possible, but observation is also possible in suburban areas. If the video camera has been newly aligned, a reference image must first be digitized, and the stars within it measured to later calculate the positions of registered meteors. With an additional MetRec program, you need about ten minutes for this. If the camera hasn't been moved since the last observation, the observation can begin in less than a minute. The camera's video signal is fed directly into the PC, which digitizes it, searches for meteors, and determines and saves all important parameters of the found meteors. After the observation, a second program helps efficiently post-process the observation, i.e., eliminating false detections usually caused by planes, birds, or insects. This takes another five minutes. In the end, you have a text file with all the data from a night, images and image sequences of the registered meteors, and database files in IMO's PosDat format. These can be read directly by other analysis programs like Radiant by Rainer Arlt to investigate the activity of known and unknown meteor showers.
MetRec was written to be easily adaptable to different cameras and tasks. The user interface and documentation are in English (there is also Japanese documentation), the program runs under Dos (or Win95/98), and many parameters can be adjusted via a configuration file for a specific PAL or NTSC camera system. The software is available for free download on the Internet for amateur astronomers.
In 2001 alone, the camera network of the Arbeitskreis Meteore e.V. (AKM) was able to record nearly 50,000 meteors in more than 6,300 hours of effective observation time. Many of these were recorded during both Leonids and Perseids, but many small meteor showers were also well covered. So far, most video data has been collected, and only a few meteor showers have been analyzed in detail. However, we plan more thorough investigations once more data is available. The data will also be made freely available to others via the Internet. We hope to gain a complete overview of the meteor streams of the northern and southern hemispheres within the next three years.
Future tasks include completing software for evaluating multi-station observations. The idea is similar: A large number of precise meteoroid orbits should be obtained throughout the year with as little manual interaction as possible. While a meteor stream's radiant can only be statistically detected with observations from one station based on a large number of stream meteors, parallel observations from two or more stations can reliably determine an orbit and radiant for each individual meteoroid. Ultimately, the software should eventually operate without additional hardware like frame grabbers and be available under current versions of Windows.