Research On Underwater Image Processing Method Based On Edge Computing Platform

Since entering the 21st century,China has officially proposed "Industry 5.0" and launched"Made in China 2025".Following the trend of industrial modernization,the robotics field,as an industry that represents a country’s modern manufacturing technology,new material research and development technology,and information control technology capabilities,has become a representative of intelligent manufacturing and industrial levels.However,the current traditional robotic arm has low flexibility and autonomy.If you don’t know the working environment or changes in the working environment,you must stop the robotic arm to avoid dangerous accidents.If you want it to work again,you need to re-adjust the robotic arm,otherwise the robotic arm will be difficult to complete the grasping task.The research topic of this thesis is the visual servo system of the 6-DOF manipulator,and the starting point is the design and research of the high-precision visual servo system.The visual servo control system is established by studying the motion model of a 6-DOF manipulator using an RGB-D camera.Realize the accurate recognition,positioning and grasping of target objects in the working environment.The main researches of this article are:1.Design the structure of a 6-DOF mechanical arm visual servo system for accurate positioning and gripping of objects,and select the appropriate equipment according to the relevant design indicators,6 Degrees of Freedom Mechanical Arm To Build a Hardware System for Visual Servo System.2.In order to study target recognition and positioning,we first explained the principle of the camera imaging model and the principle of internal and external parameter calibration.Then the principle of hand-eye calibration is analyzed,and the position-direction relationship between the camera coordinate system and the robotic arm coordinate system is obtained.Finally,the Linemod target object recognition algorithm is used to identify and identify the target object.3.Design a 6-DOF robot arm visual servo system and use the D-H method to establish and analyze a spatial description and kinematic model of a 6-DOF robot arm,and then study the principle of forward and inverse kinematics of the robotic arm and use matlb’s Robotics Toolbox robot tool Box verification,and then research on the trajectory planning of the manipulator.Finally,in view of the disadvantages of the existing RRT path planning algorithm that the search path is blind and the path is not optimal,an improved RRT path planning algorithm is improved and compared,and the comparison experiment is done through matlab simulation.4.In order to verify the operability of the design of the entire system,we will build a simulation environment based on Gazabo and simulate the visual servo system;then,based on the object recognition algorithm in Chapter 3,the common objects in life can be identified,and the pose and pose are obtained.Error analysis;then use the ROS function package easy-handeye to calibrate the robot arm "eyes out of the hand" to obtain the position and posture relationship between the robot arm base coordinate system and the camera coordinate system;finally,according to the camera feedback The converted pose information controls the robotic arm to accurately grasp the target object.