Optical and acoustic drone detection

Optical (EO) and thermal (IR)

Day cameras and thermal imagers, usually pan-tilt-zoom (PTZ). Searching the whole sky for a small drone is like looking for a needle in a haystack, so the camera is normally cued onto the target by radar or RF (slew-to-cue) and then tracks it.

Machine vision. Image recognition classifies the target – drone, bird, aircraft – and tracks it. Quality depends on training data and on-site tuning: the same model performs differently against a city skyline and an open sky.

Strengths: visual confirmation for the operator; judging type and payload (e.g. a hanging object); evidence for proceedings.

Limitations: needs line of sight; degraded by fog, rain, backlight and – for day cameras – darkness; detection, recognition and identification ranges differ; continuous monitoring needs automation, because nobody watches the sky for hours.

Acoustic detection

Microphone arrays look for the characteristic sound of rotors and motors and estimate its direction.

Strengths: passive; hears drones in radio silence; relatively low cost; works between buildings.

Limitations: short range – typically a few hundred metres; heavily degraded in noisy environments such as events, plants or busy roads.

Data protection

Camera systems can capture people. Field of view, retention and access must be governed just as for conventional CCTV.