A Fuzzy Rule Based Visual Human Tracking System for Drones

Drone sales are on the rise with each passing year and more and more people are embracing the idea of owning these devices. These unmanned aerial vehicles are now finding more practical and innovative solutions about daily problems. And also, body detection and face recognition tasks are becoming a new trend in the security authentication systems. In the light of these facts, a real time tracker algorithm is developed to track a specified human face with DJI Phantom 3 drone. After recognizing the human face, his body is detected and it is tracked. In order to improve the tracking accuracy in outdoor environments a Kalman Filter is used. When the drone misses the target, the Filter will track him for a certain time. After that a Fuzzy Logic algorithm is used to maneuver the drone towards to the human in real time. Results show that, the developed tracking system can be used to detect a human for all of the programmable drones.