| In recent years,due to the gradual maturity of UAV technology,the concept of UAV cluster and related technologies have achieved rapid development.The so-called UAV cluster refers to the cluster system composed of multiple micro and small intelligent UAVs with the same,similar or complementary functions,which can realize self-organization and collaboration through bionic means,so as to improve application efficiency and enrich mission fulfillment means.In view of the problems that the cluster recovery process of Marine fixed wing UAV is often accompanied by the interference of sea fog weather,and the uncertain landing position of each fixed wing UAV causes difficulty in positioning,this topic carries out the visual identification and autonomous recovery work of fixed wing UAV cluster based on mechanical arm in marine environment.The specific research work carried out in this thesis is as follows:First of all,a large number of relevant literature on the recovery of fixed wing UAV clusters was reviewed,and the current recycling status of Marine fixed wing UAV clusters was systematically summarized.The overall framework of the recovery system was designed,and the modeling and calibration of the visual system was studied,including camera modeling,camera calibration,and hand-eye calibration of the robotic arm.D-H method was used to establish the connecting rod coordinate system of UR5 manipulator,deduce the forward and inverse kinematics of UR5 manipulator,and Py Bullet was used for experimental verification.The trajectory planning of UR5 manipulator was carried out,including space linear interpolation,plane circular interpolation and space circular interpolation.Secondly,visual identification can be divided into two ways: adding external markers such as Ar Uco code to the fuselage of fixed wing UAV and not adding external markers.The traditional method to obtain the three-dimensional coordinates of the target based on the minimum external rectangle of the target object is not only easy to be interfered by the environment,but also has high requirements on the landing posture of the fixed wing UAV,which cannot meet the working requirements.Therefore,the design proposes an improved Ar Uco code visual recognition method.In view of the fact that fixed-wing UAVs are often interfered by sea fog in the process of maritime recovery,the existing recognition of fixedwing UAVs based on machine vision cannot accurately reflect the position relationship between robotic arm and fixed wing UAVs.Multiple experiments are conducted to compare and study the effects of different de-fogging algorithms.The visual recognition process of Ar Uco code under complex conditions is proposed as follows: The first step,the Ar Uco code recognition is fused with the dark channel prior algorithm for image defogging processing.The second step image filtering,morphological operation,contour extraction,quadrilateral screening,perspective transformation processing are carried out on the de-fogging images successively.The third step,Pn P visual positioning algorithm and other processes are used to obtain accurate 3D pose coordinates.Experiments show that this method can realize the position and pose recognition of fixed wing UAV and obtain accurate three-dimensional position and pose coordinates.Finally,based on the ROS operating system,the capture and recovery of fixed-wing UAVs is completed by controlling the UR5 robotic arm under the Move It platform,and this study proves the feasibility and effectiveness of the visual recognition and autonomous recovery system based on the robotic arm in the recovery of offshore fixed-wing UAV clusters.It provides a practical solution for the recovery of UAV swarms in marine environment,which has important engineering application value. |
