An all-sky camera is exactly what it says on the tin: a camera capable of imaging the entire sky simultaneously. This statement might sound somewhat gradious and like a major technical challenge, but in fact all-sky cameras are really just normal cameras with a special lens. This lens creates a 180 degree field of view, which results in an image that is circular but can see anything placed around it or above it (but not behind it!). By pointing the camera straight up, the entire sky is brought within the field of view of the lens, with the edge corresponding to the horizon, and the center the patch of sky directly above the camera. These cameras are perfect for studying the aurora, because no matter where it appears in the sky, you know you’ll see it!
Because of this expansive field of view all-sky cameras are a mainstay of ground based auroral measurements. In fact, there are many all-sky cameras which are setup permanently and available to be viewed online at any time. A well known one close to us in Fairbanks is the University of Alaska, Fairbanks Geophysical Institute all-sky camera at Poker Flats Research Range: https://allsky.gi.alaska.edu/.
So when planning the diagnostic suite for our aurora expedition, an all-sky camera was one of the first on the list. We had a couple of design goals for the system:
- Capable of completely autonomous operation
- Use of a modern full-frame sensor camera for maximum image quality
- Ability to maintain internal temperature at a desired level (it gets cold in Alaska!)
In order to maintain the internal temperature of the system, we knew that an insulated housing and electric heating system would be necessary to fend off the -20 degrees Fahrenheit temperatures common in Fairbanks. However, given that we wanted to use a full-frame (36 x 24mm) sensor camera, the housing would have to be large enough to accommodate this somewhat bulky camera. A meeting at the Muddy Charles was called to brainstorm for the trip and our various diagnostics, which resulted in an initial conceptual drawing:
Which was then expanded into a digital concept drawing:
With a conceptual plan in hand and the trip a go, we set off on building the housing for the all-sky camera. We chose to use a cardboard concrete form tube for the main body of the camera, plugged with 1.5 inch wooden end caps to provide lots of structure and insulation. In the bottom plate, we chiseled a notch and added a flat piece of plywood to which all of the electronics (control circuitry, batteries, and camera) could be rigidly mounted. The addition of a reinforcing rib along the back of the electronics platform ensured that it would be capable of supporting all of the internals. A few afternoons of woodworking yielded the basic form of the all-sky housing:
With the outer housing largely complete, we turned our attention to the internals. For the camera we selected the Canon EOS R, a full frame mirrorless camera with 30 megapixel resolution. For the lens, the most important part of any all-sky camera, we chose the Rokinon 8mm f/3.5 fisheye. The 8mm refers to the focal length, and those familiar with full frame focal lengths will recognize that this is an extremely short focal length, resulting in the ultra wide angle image that allows for the whole sky to be viewed simultaneously. Even so, it’s not quite short enough for the entire image circle to be visible on the camera sensor, but it’s close enough for our purposes.
But the camera itself isn’t enough, in order to make the system autonomous we need some kind of computer to control the camera and maintain the internal temperature. For this purpose we made use of the ubiquitous and affordable Raspberry Pi, a credit card sized micro computer ideally suited to low power embedded applications like this one. The Raspberry Pi can’t do the job all by itself though, and we included a second circuit board for additional electronics like a relay to control the heating element and analog to digital converters for interfacing with sensors to measure temperature etc. For the batteries, we used two power banks designed for charging cell phones, one for the camera and the other for the Raspberry Pi. They both have a capacity of 100 Watt-hours, exactly the TSA limit for carry-on Li-ion batteries.
Now all that was left was the software. In the Raspberry Pi we make use of the Python programming language and the CircuitPython library to control the heater, read temperature data from our three sensors, and trigger the camera to take a photo. The software also stores the temperature data (and other sensor data like magnetometers to be discussed in another blog post!) into text files for later analysis, and controls a set of three LEDs to indicate if the system is functioning correctly. With the software complete, a layer of paint on the housing and two new foam jackets to keep it warm, we were ready for our first proper test!
After traveling 17 hours to Fairbanks, we unpacked the system to find that it had survived the journey (much to our relief). After ensuring the batteries were fully charged and the camera storage cleared, we dropped the system in the backyard of our new home base and left it to its own devices. In the morning, we were greeted by a light dusting of snow on the camera and lots of lights flashing, a good sign that the system was still operating. We collected a total of 109 gigabytes of raw image data, comprising 3,223 images. After some processing to make the images more visually interesting we assembled them into a video, condensing our almost nine hour observing period into just two minutes of time lapse footage. While it was a cloudy night without any clearly discernible auroral activity, we were pleased that the system has demonstrated its capabilities and is certainly aurora ready! We’re looking forward to deploying the system in future nights to hopefully capture the aurora in full force.
Comments are closed, but trackbacks and pingbacks are open.