Research On Multi-camera 3D Measurement System Calibration Method And Point Cloud Fusion Method

As the most popular development direction in the field of 3d measurement,fringe projection profilometry technology has the advantages of non-contact,high precision and so on.Compared with single camera system,multi-camera measurement system has larger field of view and wider range of use.In this paper,the calibration of multi-camera 3d measurement system and the method of point cloud fusion are studied,and some new methods are proposed.The validity of the proposed methods is verified on the real system.The main work is as follows:(1)In this paper,a multi-camera system configuration scheme and global calibration method are proposed.A 3d cube calibration object is designed.The camera in each view only processes one side of the calibration object.Since the multiple sides of the calibration object are already in a unified world coordinate system,the coordinates of the feature points in the multiple sides are naturally in a unified world.The global calibration can be accomplished only by one calibration of multiple cameras at the same time.In addition,aiming at the overall error of the calibration method,a correction method is proposed,which uses the deviation between the world coordinates and the reconstructed coordinates of the feature points on calibration object to calculate correction parameters.These parameters are applied to the measurement to improve the accuracy of the whole point cloud.The results show that the proposed method is feasible and effective.(2)To solve the overlapping phenomenon of point clouds measured by multi-camera system,a point cloud fusion method is proposed to fuse overlapping points.Firstly,the position of overlapping region in the image is extracted by projectors,corresponding to the overlapping points in 3d space,using distance and normal vectors to fuse the corresponding overlapping points.For those points close to each other,an incremental weighting algorithm is used to merge the points.While for those far away from each other,the natural neighbor interpolation algorithm is used to interpolate the points.Experiments show that the proposed method can effectively deal with different situations in overlapping areas,fuse overlapping region effectively,and improve the quality of subsequent mesh.