Clouds are the bane of every amateur astronomer. Whether it is dragging a telescope out of the house to the backyard , setting an alarm for a 3:00 am observing opportunity or (especially) driving some distance to an observing site, we must always check the cloud situation. There have been many attempts over the years to use technology (remote sensing?) to help out with the obvious technique of simply looking at the sky utilizing visible light. To do this remotely would imply use of a camera and many remote observatories have such cameras in place, often all-sky cameras utilizing a fish eye lens. That still relies on a human eyeball to make the sky quality judgement. I want to explore techniques to analyze the images that can be captured by such cameras to find a criterion (or two) that will make the "judgement" automatically (perhaps turn off the alarm if need be?).
Another avenue to approach this challenge is through the physics of radiative transfer. All bodies at a temperature greater than absolute zero (which effectively means ALL bodies) transfer heat to cooler bodies via electromagnetic waves (black body radiation). The most effective wavelengths depend upon the temperature of the bodies. For most 'room' temperatures these wavelengths are in the far infrared, say ~14 u. The empty sky is very cold and high clouds are cold too - but still much warmer than the empty sky. Hence, another opportunity to detect clouds.
I remember reading years ago of someone who placed a thermistor (or some such temperature sensing device) atop a tall pole and another one at the base of the pole. The underlying physics is that, at night anyway, the ground cools off faster than the air above it if the sky is clear due to radiative transfer from the relatively warm ground to the very cold of space (hence the formation of dew, often before the Sun has fully set). These days (2021) a simple Arduino (or similar) device can easily measure the two temperatures and relay them some distance away to a suitable display. I don't know of any historical success with such a system but it would be so simple to implement that I am tempted to try it, perhaps in conjunction with some the schemes to be discussed below.