Vision Sensing And Positioning Technology For Automatic Assembly Of Leisure Umbrellas

With the widespread application of industrial robots in various industrial fields,robot technology gains continuous improvement toward high intelligent direction,including the complexity of operation tasks and the diversification of operation targets.Based on the specific task of the automatic assembly of the leisure umbrella,this paper focuses on the visual detection technology of common targets in the industrial field,and sets up the industrial robot operating platform to verifies the effectiveness of proposed method.According to the requirements of rapid calibration in industrial scenarios and synchronization during acquisition of binocular camera calibration data,a supporting image acquisition and calibration software was designed and implemented.Image preprocessing techniques such as histogram specification and mean shift filtering are determined to improve the robustness of the visual system by experimental comparison.The Hough gradient method is used to detect the horizontally placed round targets,and the distance metric between the left and right view targets is constructed.The matching problem of the corresponding targets is transformed into the optimal matching problem with weighted bipartite graphs,which can be solved using the KM algorithm.A method based on circular reprojection was used to detect the pose of the tilting umbrella cover.An improved random ellipse detection method based on connected set sampling was proposed to avoid a large number of pseudo-ellipses introduced by unconstrained random sampling points and improve the efficiency of the algorithm.The observable elliptic cone surface is standardized by the quadratic normalization corresponding to the elliptic equation,the spatial pose of the cooperative target is solved under the new coordinates and transformed into the original coordinate system,and the ambiguity in circular re-projection can be removed with the help of the geometric constraints of the workpiece.The template matching method based on SIFT feature points was used to detect irregular targets,and then the RANSAC algorithm was used to remove the mismatched points.Grabcut is used to interactively segment the target foreground from the template image,and the principal component analysis method is used to obtain the center position and main axis direction of the heterogeneous target,which provides a reference for the selection of grab points.An automatic assembly system for leisure umbrellas was built,kinematics modeling was performed on a general-purpose six-axis manipulator,and a closed inverse solution with a consistent configuration was solved using geometric methods.Finally,combined with the perception module and the execution module,the repeated assembly experiments proves the effectiveness and reliability of the system...