Usually, a microphone, or microphone array (lots of microphones), which detect the sound made by a drone and calculate a direction. More sets of microphone arrays can be used for rough triangulation.
Pros:
Medium cost, doesn’t normally locate but can provide drone direction
Cons:
Doesn’t work as well in noisy environments, very short range (max. 300-500m)
Essentially a video camera. As well as standard daylight cameras, optical sensors can be infrared or thermal imaging.
Pros:
Provides visuals on the drone and it’s (potential) payload, can record images as forensic evidence for use in eventual prosecution.
Cons:
Difficult to use for detection by itself, high false-alarm rates, mostly poor performance in dark, fog, etc.
By relying on just one detection method it can be possible for a drone to be missed. For example, when using conventional radar, it can be difficult to detect low-flying drones or distinguish drones from birds. Or if the drone is obscured by buildings or trees, an optical sensor will struggle to pick it up. By augmenting the radar and optical sensors with spectrum monitoring, the security team have a much clearer picture of any potential drone activity.
However, spectrum monitoring isn’t the answer in all scenarios. RF detection will only work when there are RF signals present. There are drones that can operate without any radio control signal and can fly using pre-programmed GPS waypoints. In these scenarios it is essential to have a comprehensive suite of sensors working together to maximize detection probability, minimize false alarms and optimize geolocation uncertainty.