Research on CAD Model Matching Algorithm and System Based on Machine Vision

In view of the problem that the local structure in the global search method of 3D model cannot meet the global features, this project intends to study the feature adjacency relationship method based on B-Rep model. By extracting B-Rep data, the feature-adjacency relationship is described. In this method, the surface, edge and other features of CAD are used to find the edge with similar features, so as to realize the segmentation of parts and CAD graphics. Comparing the adjacent matrix of the part to be recovered with that of the partitioned part can determine whether the CAD model exists or not. A general subgraph extraction method based on feature adjacency graphs is proposed to extract some features from CAD. The part to be searched is compared with the part contained in CAD, and the local similarity evaluation of the model is completed. The experiment proves that the method proposed in this paper can solve some structural problems in CAD modeling well, and the search speed can meet the demand of project search.