Research on Automatic Marking Method of 3D Model Based on Geometric Feature Similarity

Abstract

AbstractModel Based Definition (MBD) models are widely used in various manufacturing processes. The traditional automatic labeling methods still need to use pre-indexing and other preprocessing, which will reduce the efficiency and effect of labeling. Aiming at this problem, an automatic annotation method for 3D models based on similar geometric features is proposed. The Freeman chain code is used to describe the geometric shape and the contour information of the front line as the basis for geometric shape matching. The longest common subsequence is used to search and match the geometric features to determine the label object. A labeling and detection method based on case clustering is proposed, which uses case structure and information for labeling constraints and completeness detection. This paper starts from the case of car front hook cover, compares the matching results of geometric features, which turn out to be correct. The feature comparison results are very different, and the efficiency is high. Compared with traditional automatic labeling methods, the labeling time is shortened, the accuracy rate is high, and the efficiency is increased by 40%. The experimental structure shows that the method is feasible and effective, and has strong versatility, and can realize efficient automatic labeling in a short time.KeywordsAutomatic labelingMBD modelClustering algorithmFreeman Chain CodeLongest common subsequence (LCS)