Cross-granularity sheet metal part identification system and method based on machine vision technology

Abstract

The invention discloses a cross-granularity sheet metal part identification system and method based on a machine vision technology and belongs to the technical field of machine vision. For structuralappearance characteristics of sheet metal parts, a machine vision related technology is used, shape factors and rotation invariant moments of sheet metal part images are calculated to serve as coarse-grained characteristic information, sheet metal part graph contour data are extracted to serve as fine-grained characteristic information, and side face images and related characteristics of the sideface images are combined to serve as auxiliary information to construct a sheet metal part database; during detection, Euclidean distances among coarse-grained feature information are compared, similarity calculation is performed on fine-grained feature information, and cross-grained sheet metal part classification and recognition are realized; before the similarity of the fine-grained feature information is calculated, auxiliary information is compared through a template matching method and other methods, matched alternative parts are screened, the calculation complexity can be further reduced, the classification precision is ensured, and the method has good applicability to sheet metal parts with high similarity characteristics.

Description

technical field [0001] The invention belongs to the technical field of machine vision, and in particular relates to a system and method for identifying cross-granularity sheet metal parts based on machine vision technology. Background technique [0002] Sheet metal parts have extensive and important applications in the fields of automobile and aviation manufacturing. Among them, aviation sheet metal parts account for about 50% of the number of aircraft body structure sheet metal parts, which have the characteristics of many varieties, small batches, complex curved surfaces and curve shapes, various sizes, small thickness and poor rigidity. After the aviation sheet metal parts are formed and manufactured, they need to be painted in batches. After the painting process, a large number of sheet metal parts are mixed together, and it needs to be re-identified to determine the drawing number corresponding to each sheet metal part. Many sheet metal parts have the characteristics ...

AI Technical Summary

Problems solved by technology

Moreover, in the classification and identification of aviation sheet metal parts, there are problems such as the inability to design and use a unified fixture, and difficult positioning, which has caused certain difficulties in the classification work.

At present, the classification task of sheet metal parts with high similarity can only be carried out by manual comparison. This method not only has a large workload, but also has certain reliability problems, and the classification accuracy cannot be guaranteed.

[0004] At present, no recognition method for the classification of sheet metal parts has been proposed; the existing part classification method based on machine vision, because it does not combine the characteristics of aviation sheet metal parts, is directly applied to the process of classification and recognition of aviation sheet metal parts There are problems such as high computational complexity and low recognition accuracy.

Images