The UNICT-TEAM Vision Modules for the Mohamed Bin Zayed International Robotics Challenge 2020

Real-world advanced robotics applications cannot be con¬ceived without the employment of onboard visual perception. By per¬ception we refer not only to image acquisition, but more importantly to the information extraction required to carry out the robotic task. In this paper the computer vision system developed by the team of the Uni¬versity of Catania for the Mohamed Bin Zayed International Robotics Challenge 2020 is presented. The two challenges required to: 1) develop a team of drones for grasping a ball attached to another flying vehi¬cle and to pierce a set of randomly placed balloons, 2) to build a wall by adopting a mobile manipulator and a flying vehicle. Several aspects have been taken into account in order to obtain a real-time and robust system, which are crucial features in demanding situations such as the ones posed by the challenges. The experimental results achieved in the real-world setting are reported.