Research On 3D Reconstruction Method Of Workpiece Based On Point Cloud Data

Three-dimensional reconstruction has always been a hotspot in the field of computer vision.At present,3D reconstruction has been widely used in related fields such as robots,AR,and industrial manufacturing.Traditional target recognition and positioning items use two-dimensional images,but two-dimensional images are easily affected by environmental light and surface reflection of objects,thereby reducing target recognition accuracy and efficiency.If the target recognition is used with three-dimensional cloud data,the three-dimensional cloud data contains a large amount of three-dimensional information to accurately express the geometric characteristics of the target object.Therefore,3D point cloud data can be used for target recognition and positioning.Firstly,obtain the required point cloud data and preprocess the obtained point cloud data.The common point cloud data acquisition methods are studied.Through principle analysis and comparison,the GD-3d Scan grating 3D scanner produced by the 3D measurement technology equipment manufacturer,whose measurement range meets the requirements,measurement accuracy is accurate and economic and reasonable,is used to obtain the point cloud data.The obtained point cloud data contains noise points,which are removed using statistical filtering to maintain the detailed features of the workpiece point cloud.Secondly,the point cloud of a single perspective workpiece often cannot reflect the complete contour information of the workpiece.Therefore,it is necessary to register the point cloud data of multiple single perspectives workpiece to the same coordinate system,in order to obtain the complete three-dimensional information of the workpiece surface.Propose a method for point cloud registration based on contour points,which can ensure registration accuracy and efficiency.Then,3D reconstruction is performed on the obtained complete workpiece point cloud data,and greedy projection triangulation algorithm,Poisson algorithm,and moving cube algorithm are implemented.By analyzing and comparing the experimental results,the greedy projection triangulation algorithm is used to reconstruct the point cloud data of the workpiece,which can better display the detailed features of the workpiece.The reconstructed model does not have any voids,making it more accurate to reconstruct the workpiece model.At the same time,perform 3D analysis and comparison between the standard model and the workpiece reconstruction model in Geomagic Control software.3D analysis and comparison can calculate reconstruction errors.The deviation analysis results of the workpiece reconstruction model are: maximum distance: positive 0.275 mm,negative 0.230 mm,average distance: positive 0.009 mm,negative 0.005 mm,standard deviation: 0.016 mm.The final experimental results indicate that the reconstructed workpiece model belongs to the precision F level,and the accuracy of the workpiece reconstruction model is relatively high,within the allowable range of error.Finally,based on the research content of this article,the design and debugging of the experimental system are completed.The functions of point cloud preprocessing,point cloud registration,and 3D reconstruction are integrated and correlated,and the visualization of each functional module is achieved through programming,thus completing the complete process of the workpiece model.