You may have seen some hardware equipment poking out from the bottoms of some of the PhenoCam images — poles and panels.
Why are they there? And what are they for? Early in the days of the PhenoCam network, we were coming up with a way to convert images into meaningful numbers that could be compared over time. One of the key questions was how to measure the greenness of an image. This was particularly tricky, because all sorts of things can affect how “green” an image looks. The angle of sunlight, for example, or haze between the camera and the foliage, alter the way colors show up in the images.
An early thought was to have a way to correct for variations in lighting and air quality. If we had a known color in the image, then we could adjust the rest of the colors using it. That’s where the panels come in. Panels like the one above have a known color and we can adjust the balance of the rest of the image based on it.
There was only one problem: the angle of sunlight and air quality might also affect the way the panel appears in images. We handled the air quality issue by putting the panels close to the camera, so that any haze would be minimized relative to the distance to the vegetation. The sunlight angle problem could also be handled by making the panels out of a material called Spectralon. This material diffuses light such that its apparent brightness is the same no matter from what angle you look at it. But it’s expensive. Very expensive. Spectralon panels did go up at a few sites — for example in the image above and the one below.
But they were too expensive to put everywhere. So an effort was made to see if regular old painted panels would work well enough. You’ll see some of those in the images, too. Some are just white, and others are multi-colored.
It turns out that painted panels don’t work as well as Spectralon. In the end, we used a greenness metric that was fairly insensitive to lighting conditions, and so didn’t need to use the panels to calibrate the images after all. That’s why you see them at only a few sites.
We did still use the panels for other things, though. By using the panels as a reference, for example, we were able to show that the sensors in the cameras don’t degrade over time — at least not in the decade or so that the oldest cameras have been up. That’s important because it means we can compare the signals we get today to those from many years ago, without having to make adjustments to the data.