Research On Automatic Pipeline Alignment Technology Based On Machine Vision

Pipeline alignment is a crucial link in the pipeline alignment and welding process.Precise alignment can reduce the occurrence of misalignment and ensure welding quality.In order to solve the pose adjustment of automatic pipeline alignment,a method of using machine vision technology to guide the attitude adjustment mechanism to perform pipeline alignment is proposed.In this paper,the pipeline automatic centering control process is designed,and the four-degree-of-freedom pose adjustment platform and the software development of the host computer are completed.The control system adopts the main control scheme of industrial computer and motion control card.In order to meet the visual measurement requirements of the pose deviation between pipelines,this paper focuses on solving the following three key technical problems:(1)Visual measurement system calibration: Firstly,the perspective projection imaging model of industrial camera is analyzed,and a linear structured light measurement model is established.Secondly,the camera calibration process based on Halcon software is studied,and accurate camera internal and external parameters can be obtained by this method.Finally,the calibration method of the line structured light plane based on the calibration plate is studied.The method first measures the threedimensional coordinates of the laser stripe center point in the camera coordinates under different calibration plate poses,and then uses the orthogonal regression method to fit a plane to get the parameters of the laser plane.(2)Rotation parameter calculation: This method constructs a monocular vision measurement system,and under the premise that the internal parameters of the camera and the pipe size are known,according to the pose solution model of the circular feature,the closed-form solution of the circular pose of the nozzle feature is reversely calculated by using the ellipse feature of the nozzle image.In order to solve the ambiguity of the pose solution,the true pose of the feature circle of the nozzle is obtained by using the constraint that the normal vectors of the inside and outside feature circles of the nozzle are parallel to each other.In the process of ellipse feature extraction of nozzle image,an improved method of least squares ellipse detection based on arc combination is proposed,which can quickly and accurately detect ellipse features in images with different poses under natural lighting conditions.(3)Calculation of translational parameters: This method is used to measure the three-dimensional topography of the groove by constructing a line laser vision measurement system.For the translation deviation in the X and Z axis directions,the three-dimensional coordinates of the laser stripe feature points in the camera coordinate system are obtained for calculation;for the translation deviation in the Y axis direction,the calculation is performed by detecting the centerline of the cylindrical surface of the pipe.After experimental verification,this method can be employed to guide the pose adjustment platform to complete the automatic centering of the steel pipeline with a diameter of 80 mm and a length of 200 mm.The measurement error of angle deviation is less than 0.5°,and the measurement error of translation deviation is less than 0.3 mm.This technology can provide effective technical support for the research and development of automatic pipeline alignment device in the future research.