Plots Explained
GMN plots and images explained by Damir Šegon There are more than 1,000 Global Meteor Network (GMN) cameras sharing the web pages, powered by GMN. The GMN Weblog features their image stacks and plots, and summarizes the previous night's Raspberry Meteor System (RMS) camera results.
The GMN weblog page is here: https://globalmeteornetwork.org/weblog/. Results also are available on the GMN status page: https://globalmeteornetwork.org/status. This document is a guide for typical RMS processing.
Some of these stacks/plots are easy to understand at a glance, while others require more explanation. For image stacks, CMNbinViewer software is a valuable tool that you can run on the Pi or PC. There are publications listed at the bottom of the main wiki page about RMS, which may help you understand how best to capture and detect events.
Detected stack
A stack of detected images is an event that was recognized as a meteor. A detected image is the difference between MaxPixel and AveragePixel images, which is calculated in a procedure that helps suppress slow-moving clouds and most static noise. Fast moving clouds may show up on the detected stack as a set of ‘waves’ that form what appears to be an abstract art painting.
After the real-time analysis of images, several machine learning algorithms have been trained to reduce false detections from clouds flying creatures, such as bugs, bats, seagulls, and owls, or other flying entities, such as airplanes, drones, and satellites.
In spite of false event rejection filters some false positives may remain. Also note that if a plane or other false event occurs while a meteor is detected, both the meteor and the false detection will be captured in the same image. As a result, the note in the bottom right corner, Detected meteors, should be understood as an estimate. For stations with good astronomical viewing conditions, this number will be very close to the number of detected meteors.
Captured stack
In contrast to the Detected stack, a captured stack contains images from the entire night, some of which also are represented in the Detected stack.
Detected thumbnails
Similar to a Captured thumbnail, a ‘’Detected thumbnail’’ represents a matrix of only MaxPixel images that contain detections. Time stamps on these images specify a start point you can use to search for an interesting event. Unlike images in a Captured stack, Detected thumbnails are not stacked, which means that each image represents only one FF file.
Captured thumbnails
This image is built from a matrix of MaxPixel images from the entire night, and it includes time stamps in the upper left row. Each image is actually a MaxPixel stack of MaxPixel images, which are the result of combining 5 FF*.fits files.
The time difference between two neighboring time stamps should be 10.24 x 5 = 51.2 seconds. If this difference is higher than 51.2 seconds, your RMS system dropped frames and you should check for problems.
Because MaxPixel values are stacked, the sky sometimes is too bright, which causes images to become saturated (completely white). This result can be is a combination of moonlight and high humidity or lightning that flashed in your field of view (FOV), illuminating the clouds.
Radiants
‘’Radiants’’ are plots of meteors and active shower radiant positions. The size of the circle for a radiant depends on the meteor shower, and counts are provided for meteors associated with an active meteor shower.
A radiant association is based on single station observations, which may be wrong when combined with other stations. Usually, counts are highly accurate for a significant event or major shower activity. In the lower part of the plot, the vertical dashed red line represents the current position of the Earth relative to solar longitude and active radiants on the date of observation.
Timestamp Intervals
This plot shows relative stability of frame recording throughout the course of the night.
Deaveraged field sums
This plot shows variations between pixel value sums from each FF file throughout the night. Because these values are deaveraged, any significant event peak should be obvious in the plot.
Peak field sums
This plot shows the sum of intensity values from all pixels in each frame, from each FF file throughout the night. Average values from these sums are plotted in black and peak values are plotted in red. As before, each significant event in the camera field of view is shown as a peak.
Astrometry report
This plot shows the results of an image that was refitted during the night, with deviations from original calibration shown as yellow lines. The length of each line is 100x the distance from ideal calibration and the line orientation shows the direction of the deviation.
If there is a systemic error, lines display either as concentrically oriented or deviated in a single direction. Ideally, you want to see a lot of short lines that are oriented in multiple directions.
Calibration variation
In early stages of the GMN while checking astrometry, it was discovered that cameras can significantly change their orientation during the course of the night. Why this situation occurs is beyond the scope of this document, but it is important to know that this movement requires you to check the astrometry fit for each FF file that contains a detection.
The results of a recheck are shown in this plot and the time from the first FF file is color-coded. The figure shows the difference between the refitted FOV and the reference FOV center in arcminutes, usually on the X-axis. It also shows the angle of rotation between the refitted FOV and the reference FOV.
If you study this plot for your camera through the night, you may be able to identify and fix problems related to the camera mount. On average, cameras move 5 to 6 arcminutes overnight, but there are documented cases of minor movements of only 2' and much larger movements of 15' or more.
Photometry report
This plot shows the photometry fit for stars extracted from an FF file of observations. Two lines represent the fit for the recent night (red) and a reference calibration (gray).
To calculate the limiting magnitude of a camera, subtract about 4 magnitudes from the photometry fit. For most 3.6mm cameras, this number is 10 or a bit more because these cameras have a limiting magnitude of about 6.
Photometric offset variation
Photometric offset over time
Flux total observing tim
Graphic plot Matched vs. Predicted stars throughout the night.
Masked Flat
Flat after overlaying mask.bmp
Observation Summary
Observation summary data of hardware and data recording characteristics.
Meteor Shower Flux
Meteor shower flux charts, if specific showers are detected.
TimeLapse video
This TimeLapse.mp4 video contains many images from the previous night, but not all images. This video is not valuable for scientific applications, but it is useful to check dynamic weather conditions throughout the night.
Acknowledgements
Plots and images in this document are courtesy of Danijel Reponj, Observatory Apollo, Croatia, GMN station HR000S. His camera, Cyclops II, is a replacement for the first analog camera, which ran 24/7 for 8 years!