Autonomous Quadcopter Flight System with Object Tracking

Abstract

Quadcopter is an aircraft similar to helicopter, but lifting and propelling by four propellers. It is a helpful tool for aerial photography and real-time monitoring. The aims of this project is to develop an autonomous system for quadcopters which can accomplish some tasks automatically. The system has two functions, orange ball tracking and selfie function. The selfie function was developed with the help of face detection. The system was implemented on Crazyflie and AR.Drone 2.0. The camera image was fed into computer for process. By analyzing series of images, it is possible to determine an object’s position and movement and hence generate commands to allow a quadcopter to perform suitable action to keep tracking objects. However, due to the weight and power limitation, Crazyflie did not achieve a good performance in this project. On the contrary, AR.Drone had demonstrated it has a really high ability for objects tracking. The results of this project show quadcopters have the ability to track objects accurately without human interruption and have potential to execute more intelligent tasks. With this ability, quadcopter can be used on some austere environment to perform dangerous tasks or developed more applications with the help of camera and objects identification.