Development Of Structural Part Flatness Measuring Robot Based On Binocular Vision

With the development of urban rail transportation,the subway travel mode has become the main public travel mode because of its comfortable,punctual and convenient characteristics.The market demand for subway related facilities has increased sharply,and the production quality requirements have become increasingly strict.The measurement of flatness has become a key process that affects the production efficiency of shield door facilities.Therefore,the development of a robot to measure the flatness of structural parts with T-shaped lower support is of great importance for improving the production efficiency and quality management of subway screen door enterprises,and it is also of great significance for the promotion and reference of flatness measurement of similar structural parts.The flatness solution process,that is,the depth information solution process of the surface feature points to be measured.In recent years,structured light and binocular vision have been successfully applied to the field of 3D ranging.Based on this,this paper conducts an in-depth study on the development of a flatness measurement method and measurement robot based on structural light and binocular vision for subway shield door structural parts.Firstly,the hardware design of imaging system and binocular camera fixture is completed based on the requirements of flatness detection technology,and the calibration of binocular camera and robot hand-eye is carried out;the research focuses on the structured light grid intersection localization algorithm based on the confidence degree.The three-dimensional coordinates of feature points are measured based on the binocular ranging model,and the least squares method is applied to solve for the flatness of structural parts.The effectiveness and superiority of the planimetric measurement method in this paper are verified by comparing with the planimetric measurement method based on laser displacement sensors,and the efficiency and cost of developing a measurement robot are verified by comparing with the manual planimetric measurement efficiency and cost.The main research work of this paper is as follows.(1)Based on the analysis of the flatness inspection requirements of T-shaped lower support,a scheme design of a structural component flatness measurement robot system based on structural light binocular vision is carried out.The imaging system was built and the binocular vision system was integrated into the robot to realize the automated imaging of the surface to be measured on the T-shaped lower support,and the selection of robot,camera and structured light,as well as the binocular camera binocular calibration and robot hand-eye calibration were completed.(2)A confidence-based algorithm for structured light grid intersection localization is investigated.Because of the similarity between the structured light grid intersection and the checkerboard grid intersection in terms of geometric features,the Harris angle point feature extraction algorithm is used to extract the network intersection first,and the experimental results show that this method is not suitable for this application due to the poor robustness.The confidence-based grid intersection localization algorithm is designed based on Geiger’s idea of calculating the confidence degree in the checkerboard grid corner point extraction algorithm for grid structured light geometric features.The algorithm achieves grid intersection localization through four steps: image pre-processing,confidence calculation,non-maximal value suppression,and intensity distribution-based rechecking.By analyzing the grid intersection extraction results of 40 structural members to be tested,the effectiveness and good robustness of the confidence-based grid intersection localization algorithm applied to T-shaped lower support are verified.(3)The algorithm for calculating the flatness of structural members is studied,and the flatness measurement algorithm proposed in this paper is analyzed and evaluated.To further improve the positioning accuracy of the grid intersection,a sub-pixel positioning algorithm based on Gaussian fitting is designed based on the feature that the grayscale values of the pixel points in the neighborhood of the grid intersection are centrally distributed,and the grid intersection is sub-pixel positioned by two steps of coarse positioning and fine positioning in turn;then the left and right eye grid intersection matching is performed based on the polar line constraint,and based on the 2D image coordinates of the grid intersection with binocular calibration and hand-eye calibration Based on the results,the 3D coordinates of the grid intersection are calculated,and finally the planarity of the structure is solved based on the least squares method.The results are compared with those of the laser distance measurement sensor and the Hoff transform grid intersection positioning algorithm to verify the effectiveness and efficiency of the flatness measurement algorithm proposed in this paper.(4)The development and testing of the robot based on binocular vision for measuring the flatness of structural parts of subway shield doors,and the analysis and evaluation of the robot in terms of anti-interference capability,efficiency,and economy,respectively.The experiment verifies that the robot motion and ambient light have minimal influence on the robot flatness measurement results and have good stability in the factory environment;the comparison with the manual flatness measurement efficiency verifies the high efficiency of the flatness measurement robot measurement;finally,based on the cost calculation,it verifies that the developed flatness measurement robot has a high return on investment.