| In today’s era of continuous scientific and technological progress,mechanical arms are widely used in all walks of life.However,some heavy-duty mechanical arms are always restricted by environmental factors when carrying out lifting tasks,which will lead to low success rate of handling and long planning time.With the development of machine vision,artificial intelligence and other technologies,the combination of vision and mechanical arm enables the mechanical arm to have the ability of autonomous visual positioning and detection,which is of great significance to improve the detection ability of the mechanical arm and achieve flexible assembly,adaptive grasping and precise positioning of the mechanical arm.In this paper,a dual camera non-common field of view pose measurement method is studied based on the automatic and punctual loading of cargo by lifting manipulator.The main research contents are as follows:1.Expounds the working background of the visual pose measurement system,establishes the camera imaging coordinate system,and analyzes the principle and method of camera calibration.Based on Zhang Zhengyou’s “planar checkerboard” calibration method,the dual-camera calibration was carried out,and the camera internal parameters were obtained through the left and right camera calibration experiment.2.Research on image feature extraction and matching and center point location algorithm.Firstly,the image collected by the camera is preprocessed,and the median filtering algorithm is selected to smooth the image.Based on the extreme point features of the image,SIFT algorithm is used to extract and match the image features,and RANSAC algorithm is used to optimize the feature extraction algorithm.After matching the image,Canny operator is used for edge detection.Because the cooperative target is a circular target,the least square fitting algorithm is finally used to obtain the center point of the cooperative target.3.In view of the working background of automatic alignment of lifting goods carried by lifting manipulator arm,based on the non-common field of view system architecture of dual-camera pose measurement system,two cameras are installed on the lifting manipulator spreader and two cooperative targets are set at the bottom of the vehicle carrying box,and the pose solution model is derived when the spreader is parallel to the bottom of the vehicle carrying box.The correctness of the calculation model is verified by MATLAB and physical simulation.The experimental platform of dual-camera was built and the pose measurement experiment was carried out.The experimental results show that the proposed method has practical application value.Aiming at the working background of automatic alignment of lifting goods carried by lifting manipulator arm,based on the non-common field of view system architecture of dualcamera pose measurement system,this paper installed two cameras on the lifting manipulator spreader and set two cooperative targets at the bottom of the vehicle carrying box,and deduced the pose solution model when the spreader was parallel to the bottom of the vehicle carrying box.The correctness of the calculation model is verified by MATLAB and physical simulation. |
