Research On Adaptive 3D Measuring And Damage Extraction Method For Remanufacturing Components

Remanufacturing is one of the best ways for the recycling of electromechanical products,which can provide key technical support for circular economy,and has become one of the most promising technical fields in the world.Remanufacturing components are generally large in overall size,while damaged parts are tiny,and vary in number and locations.This makes it difficult to measure remanufacturing components and reconstruct models of damaged parts,and eventually reduces the overall efficiency of remanufacturing greatly.This thesis takes the adaptive 3D measuring and damage extraction method for remanufacturing components as the research object.A normalized skeleton model construction method for 3D models was researched.In order to rapidly extract highprecision data of damaged parts,a skeleton-based overall adaptive rough scanning method for remanufacturing components and an adaptive fine scanning method for damaged parts were researched.Corresponding experimental cases were carried out.The main research work is as follows:(1)To resolve the problem that skeletons extracted by existing methods cannot express topographical and positional information accurately,a normalized skeleton model construction method for 3D models was researched.Skeletons of mesh models and point cloud models were extracted on the basis of a skeleton extraction algorithm via Laplace contraction.Mesh models were segmented on the basis of a skeleton-based 3D model segmentation algorithm.Normalized skeleton models were constructed according to the results of skeleton extraction and model segmentation,which contain accurate topographical and positional information of standard mesh models.(2)To resolve the contradiction of work efficiency and data acquisition accuracy in the 3D measurement of remanufacturing components,a skeleton-based overall adaptive rough scanning method was researched.The overall adaptive scanning route was generated according to the normalized skeleton model.The overall rough scanning accuracy was settled according to the minimum damage scale.Experiments were performed by using a linear structure-light binocular stereo visual 3D measuring robot,rough point cloud data of every branch of scanned damaged component was collected exhaustively in a small data volume.(3)Aiming at the target of rapidly damage localization of remanufacturing components,a point cloud preprocessing and damage localization method based on contour models was researched.The standard mesh model and its corresponding point cloud model were registered rapidly by matching their skeletons.The contour model of the standard mesh model was built according to the normalized skeleton model.Noisy points and redundant points were removed based on standard contour lines.Rough damage points were extracted based on the standard contour model,and then locations and abbreviated damage boundaries were figured out.(4)Aiming at the requirements of high-precision data acquisition of damaged parts of remanufacturing components,an adaptive fine scanning and model reconstruction method was researched.Adaptive fine scanning strategy for damaged parts was researched on the basis of damage locations and abbreviated damage boundaries,fine scanning regions were figured out,and adaptive fine scanning routes and accuracy were determined.Corresponding scanning experiments were performed,and data volumes of the proposed adaptive scanning method and the traditional scanning method were compared intuitively.Noisy points in fine scanning point clouds were removed based on the contour model.Fine point clouds were fused with the overall rough point cloud,high-precision surface data of damaged parts was extracted and mesh models were reconstructed.