Visual Detection of Drones


In our series of blogs on Drone Detection Techniques, passive detection was briefly discussed. Before we move on to a multi sensor drone detection system, it is important to touch upon visual detection as a technique to detect drones which helps in tracking and identification of drones as well.

Passive detection systems depend on detecting UAVs using sensors that do not actively  radiate energy.

Electro-Optical/ Infra Red (EO/IR) cameras and Thermal Imaging (TI)  sensors, fall in this category of passive detectors. As discussed in our earlier blogs, Autonomous drones cannot be detected by RF Detectors as they do not rely on RF signals for navigation. Radar detection ranges of micro and nano drones are limited due to the small size of the drone. Even in the event of detection, there still remains the problem of false alarms as they can be mistaken for medium sized birds such as eagles and kites. In such situations, EO/IR and TI sensors are an invaluable means of detection. PTZ or Pan-Tilt Cameras can be used to get visuals on the detected drone, and confirm a drone threat. They have the added advantage  of better accuracy in location fixing as compared to RF and Radar detectors, and hence permit coupling with hard kill means such as Air Defence Guns to bring down hostile drones. Since cameras have an inherent video recording facility, they provide invaluable information of drone intrusions into sensitive areas.


An optical sensor works by capturing the ambient light that reflects off the drone and reproducing the image. The effectiveness of the optical sensor depends on the size of its aperture and the focal length of the sensor lens. Normally, it is a trade off between these two parameters since a wider aperture implies a larger field of view but reduced magnification since its focal length would be smaller than in the case of lens with narrower apertures. The smaller field of view associated with high magnification implies more time is needed to scan the horizon for intruders. Normal visual detection ranges achievable are between 1.5 – 2 Kms in good daylight conditions.


In conditions of poor visibility such as fog or night time, detecting the heat emitted by the drone is a better option. IR cameras do precisely that by capturing the energy radiated by the drone in the near-IR portion of the spectrum. Since this part of the spectrum is very close to the visual portion of the spectrum captured by the EO sensor, EO and IR sensors are frequently clubbed together. Indeed normal commercial cameras have a specific filter to block IR. In our application, though the IR energy is specifically detected. TI cameras are more complex although they too detect heat energy but in parts of the mid and far IR spectrum. They work by determining the temperature difference between the drone and its ambient environment and produce a heat image rather than an optical image.



EO/IR and TI sensors offer the possibility of image processing and machine learning to help in classification of the drone.



Please stay tuned to the next blog on multi sensor drone detection system.