...Also random interesting things like optical fiber-based sensors, where an optical structure at one end of the fiber exhibits optical characteristics that change in response to an external condition like temperature or strain, which can be read by a light source+detector at the other end of the fiber...
You can also use the optical fibre itself as a sensor. Over the years, I've worked developing systems using laser light pulses, launched into one end of a fibre, which are backscattered as they interact with the molecular scale structure of the glass, returning to the launch site as a delayed 'echo', much like a radar or lidar system. Except that the backscatter signal is continuous, being generated at each point in the fibre as the laser pulse reaches it. Simple Rayleigh scattering returns light at the same wavelength as the input pulse, and forms the basis of the OTDR (Optical Time-Domain Reflectometry) method of measuring attenuation in fibres. However, if you use an extremely narrow linewidth laser (~kHz) and coherently detect the backscatter, using a frequency-shifted copy of the laser output as a local oscillator, you can measure microscopic
displacements of each point along the fibre. This lets you use the fibre as a distributed vibration sensor or microphone.
Raman scattering is a non-linear interaction which causes the wavelength of the backscattered light to be different from that of the incident laser pulse. The Raman scattering process is temperature-sensitive, so you can make a distributed thermometer or temperature sensor based on the intensity of Raman backscatter. It isn't easy, as the amount of Raman-scattered light is very much less than the incident pulse, and there are all sorts of complications arising from the variations in fibre attenuation and group velocity at the different wavelengths involved. However, the wavelength shift is large enough to enable the Raman backscatter to be easily separated from the Rayleigh signal using optical filters.
Brillouin scattering is another non-linear process, also sensitive to temperature but sensitive to strain as well. The level of Brillouin scattering is intrinsically much larger than that of Raman scattering (unless the fibre has been specifically designed to suppress it - as some high-performance communication fibres now are) but the wavelength shift is much smaller. Optical filtering is no longer possible, so coherent detection using an optical local oscillator, followed by standard microwave IF technique, is required to recover useful measurements.
These sensors have many applications, mostly measuring long, thin things. Like oil wells, gas pipelines, security fences, railways, and even optical communication cables themselves (using either a spare fibre or 'spare' wavelengths in a traffic-carrying fibre). To extend the range, it is possible in some cases to use intermediate optical amplifiers (EDFAs) which are themselves 'pumped' (i.e. powered) by light sent from the measuring instrument.
It's an involved and fascinating field. The technical problems are non-trivial, and as well as photonics 'proper', can require expertise in analogue electronics, digital signal processing, measurement physics, and materials science & chemistry.
For more information, I recommend finding a copy of
An Introduction to Distributed Optical Fibre Sensors by my former colleague Arthur H. Hartog, published by CRC Press Taylor & Francis Group, ISBN 978-1-4822-5957-5