Research On Positioning Technology Of Spherical Pipe Inspection Robot

Pipeline robots are important tools for detecting pipelines and clearing obstacles.The movement of traditional pipeline robots within the pipeline is affected by the path,and the action is limited,which affects the inspection results.In this thesis,a spherical pipeline inspection robot is designed to solve the problem of robots moving freely in complex paths.The electromagnetic wave technology,image processing algorithm and sensor positioning technology are used to realize the external positioning of the robot inside the pipeline and the detection and location of the inner wall defects of the pipeline..The main research contents of this thesis include.(1)Hardware design of spherical pipeline inspection robot.The 3D stereo camera is selected,and the video stream is captured with the illumination light band,and the low frequency signal transmitting device,the displacement recording device and the power supply source are designed.The experimental pipeline of the spherical robot by the fluid thrust suspension movement inspection was designed.(2)The external positioning system of the extremely low frequency electromagnetic wave spherical inspection robot is designed.The 22 Hz low-frequency signal transmitting and receiving magnetic field strength model is established by simulation,and the coil parameters are determined.The signal transmitting coil and the receiving sensor array are constructed.Using the quasi-Newton particle swarm optimization algorithm,the optimal solution of the equations is obtained by iterative convergence,and the robot is obtained.Position coordinates.The error of the external positioning algorithm of this experiment is less than 22 cm.(3)Image processing techniques are used to extract pipe defects containing defects and welds.The image pre-processing technique is used to filter the noise and blur recovery of the captured image,extract the HOG features of the pre-processed image,and train the SVM classifier to classify the image of the inner wall of the pipeline.The image classification accuracy can reach 76%.(4)Using the nine-axis sensor attitude data and defect recognition results to achieve internal positioning of the defects in the tube.The nine-axis sensor data is obtained by using the attitude calculation algorithm to obtain the gravity component,and the displacement data is obtained by integrating the acceleration data to realize the recovery of the robot motion track,and the time correspondence is related to the position coordinates of the defect in the tube.