Structure Of Underwater Culvert Dredging Robot And Its Research On Visual Positioning Technology

At present,the South-to-North Water Diversion Project has achieved relatively satisfactory results,which has solved the problem of lack of freshwater resources in parts of northern my country to a great extent.However,in the process of water conveyance,silt is produced at each culvert mouth of the water conveyance channel.The long-term accumulation of silt will cause environmental pollution and affect people’s daily life.At the same time,there are obstacles on the road surface in the underwater culvert that hinder the equipment from moving.,And it is difficult to clean up manually.Aiming at this background,this paper designs an intelligent underwater dredging robot that can clear obstacles on its own in order to solve the above problems.Combined with the actual working conditions,the overall design scheme of a new type of intelligent dredging robot applied to the silt cleaning of underwater inverted siphon culverts is given,including the design scheme of each part of the mechanical structure,the control system scheme,and the camera configuration scheme for assisting the positioning of the robot arm.Obstacle clearing method process for visual positioning.Aiming at the characteristics of the dredging robot working underwater,the water resistance simulation analysis is carried out by simulating its working environment,and the water resistance of the entire equipment under the working environment with the highest expected flow rate is obtained,and the robot is verified by calculation at the bottom of the culvert with silt.Walking ability;In addition,it focuses on the kinematics and dynamics analysis of the obstacle removal manipulator.First,the DH parameter method is used to establish the kinematics model of the manipulator.Based on this model,the forward and inverse kinematics analysis is performed,and the random point method is simulated in Matlab.The working space range of the manipulator is determined,and then the dynamics model of the manipulator is established based on the dynamic equations of Newton-Euler method,and the dynamic parameters such as torque and angular velocity are solved.Finally,the underwater working environment of the manipulator is taken into consideration.The influence of resistance on the torque of each joint of the manipulator.First,through Fluent simulation,the resistance coefficient of the manipulator working under water is obtained.The obtained water resistance coefficient is combined with the Morsion equation to establish a hydrodynamic model and analyze its mechanical characteristics in the presence of water resistance.When performing obstacle clearance work,the movement of the robotic arm needs to rely on visual positioning as a guide.First,an imaging model of the camera is established based on the currently generally applicable small hole imaging model,and the conversion relationship between the coordinate systems related to the model is given,and the imaging is combined The model explains the method of solving the three-dimensional coordinates of the spatial point based on the binocular vision model;then the blanking point model is introduced to solve the problem of the poor accuracy of the internal parameter initial value of the traditional Zhang’s calibration method and used as the method to obtain the initial value of the internal parameter;for the water in the culvert Under severely insufficient illumination,the overall image is dark and accompanied by complex noise,combined with mainstream image enhancement methods,a hybrid and adaptive image enhancement method is proposed;finally,considering the requirements of positioning accuracy and work efficiency,a Census-based image enhancement method is adopted.The transformed semi-global stereo matching algorithm performs stereo matching on the image to obtain the disparity,and obtains the depth information based on the disparityIn order to verify the feasibility of the improved and optimized algorithm,an experimental platform that simulates the underwater working environment was built,and the improved calibration method and stereo matching algorithm were verified by experiments.The results showed that when the number of samples is greater than 20 calibration images,the improved The calibration method has very little error from the traditional Zhang’s calibration method,and the improved calibration method has faster error convergence speed and higher calibration accuracy under the condition of a small number of samples.The optimized stereo matching algorithm can also be obtained within 500 ms Coordinate results with relative error less than 3%.