Research On Working Space And Visual Technologies Of Underwater Rigid-flexible Manipulator

At present,underwater manipulators have been widely used in the fields of ocean exploration and resource exploitation,but the development of related technology is not comprehensive enough.For example,there is less analysis on the working space of underwater manipulators,which cannot guarantee the accuracy and safety of operations under different working conditions;The development of visual algorithms is not thorough,lacks the detection function of visual window pollutants;In the case of visual pollution,the cleaning technology of pollutants is not proper,which is liable to affect the accuracy and safety of operations.In view of the above problems,the following researches are carried out in this paper:Firstly,the working space of a kind of underwater rigid-flexible manipulator is solved and its impact factors are analyzed.Based on the hypothesis of piecewise constant curvature,the kinematics model of the rigid-flexible manipulator is established,and forward kinematics and working space of the manipulator are solved based on the principle of minimum strain energy.Then the working spaces under different input conditions are calculated and the influence of the impact factors are analyzed.The impact factors of the manipulator’s working space are evaluated based on the Analytic Hierarchy Process and the weight of each impact factor is determined.Secondly,an underwater target and pollutant detection algorithm is established based on deep learning,an underwater ranging algorithm is established based on deep learning,and the two are combined to achieve visual recognition and location.The underwater target detection dataset is established,and the underwater target and pollutant detection algorithm is established based on the YOLOv5 target detection algorithm.The checkerboard calibration method is used to calibrate the underwater binocular camera and the underwater ranging algorithm is established based on binocular vision.The depth information of the environment can be obtained and 3D reconstruction can be carried out.The underwater visual recognition and location is realized by combining the target detection algorithm with the ranging algorithm.Thirdly,a pollutant cleaning device is designed based on the water-jet technology,and a pollutant cleaning algorithm is established based on the Case-based Reasoning theory.The water-jet cleaning module of the cleaning device is designed based on the water-jet technology,and the pollutant cleaning area corresponding to each jet nozzle is divided.Subsequently,a wiper cleaning module is designed as the spare cleaning module of the cleaning device,and a silicone cleaning brush is made by means of silicone mould casting.The pollutant cleaning case base is established and the pollutant cleaning algorithm is designed based on the Case-based Reasoning theory.Finally,the experimental research on the visual window technology of the underwater rigid-flexible manipulator is carried out,and the device and algorithm designed in this paper are verified.Based on the laboratory hardware facilities,the overall control platform is built,and the software systems are designed and developed.The underwater target detection and ranging experiment is used to verify the visual recognition and location algorithms,and the pollutant cleaning device and cleaning algorithm are verified by the visual window pollutant cleaning experiment.