Quadcopters are highly manoeuvrable and can provide an effective means for an agile positioning of sensors such as cameras (flying cameras). In this talk I will present how visual sensing can allow flying cameras to fly in formations without the need of external positioning systems such as GPS. I will address the problem of formations following a moving (ground) target, and explain a distributed decision-making process performed via sensing and sharing of information among neighbouring flying cameras. Given formations of up to 12 flying cameras, I will show how the target-following performance varies when the visual sensing is used as opposed to a GPS-only-based positioning.
Formations of flying cameras to film a moving target
Fabio Poiesi is a postdoc in the Centre for Intelligent Sensing at Queen Mary University of London. He received his PhD degree from the same university in March 2014, supervised by Prof. Andrea Cavallaro. He published in journals such as IEEE Trans. on Image Processing, IEEE Trans. on Circuits and Systems for Video Technology, Computer Vision and Image Understanding, and in conferences such as IROS, ICIP and AVSS.