Monocular Vision Ranging Method For Circular Welded Seam On Inner Surface Of Cylindrical Pipeline Based On Optical Imaging Model

The positioning of the weld seam is a basic and important technology for the intelligentization of pipeline flaw detection and inspection robots.However,due to the lack of points and linear characteristics of the circular weld in the pipeline,the automatic locating of the circular weld seam in the pipeline by the monocular vision robot is still a challenging research job.This paper studies how to use the prior knowledge of the contour shape of the circular welded seam in the pipeline to assist the locating of the annular weld.This paper(1)constructs a mathematical model for the positioning of the circular weld seam with no movement of the robot,that is,the geometric imaging model of the weld seam of the pipeline non-destructive inspection robot when there is no swing and is facing the circular welded seam,which is obtained from the established imaging model.The shape of the circular weld in the image plane of the robot camera is an ellipse when the aspect ratio of the camera imaging pixel is not 1.The frame difference method is used to segment the welding seam image target and the parameters of the welding seam contour ellipse model are obtained by the Hough transform which improves the calculation efficiency.(2)Based on the obtained circular weld geometry imaging model,the relationship between the distance between the robot and the weld and the profile parameters of the weld is established,and the pixel aspect ratio and focal length are considered when the robot detects the movement of the circumferential weld.In the process,the constraints remain unchanged.The least squares solution method is used to suppress the influence of the image segmentation error of the circumferential weld,and the distance between the robot and the circumferential weld in the physical space is solved.(3)On the basis of the no-swing interference model,the problem of the robot movement with left and right swings interfering with the positioning of the circular weld is studied.Since the swing angle of the motion cannot be measured more accurately,the problem is to complete the measurement of the distance from the girth weld to the robot.It involves the solution of the deflection angle of the robot movement(such as the deflection angle of the camera),which forms an obvious SLAM problem.In this paper,based on the establishment of an imaging model of a circular weld that interferes with the movement of the robot,combined with the speed information of the robot,a control equation for the distance between the positioning of the circular weld and the detection robot for this special SLAM problem is established,and the problem of solving the control equation is transformed into an optimization problem that uses a loss function to minimize it.In view of the high nonlinearity of the optimization problem,this paper uses a genetic algorithm to solve the optimization problem,and realizes a positive loop under the interference of the left and right swing of the robot movement.Solve the weld location.(4)Using the synthetic video of the robot to detect the circular weld in the cylindrical pipe and the video captured in actual work,the image detection of the proposed circular weld,the extraction of contour model parameters and the weld in three-dimensional space Positioning of the distance to the robot.Experimental results show that the positioning method proposed in this paper has achieved good results.