Research On The Detection Method Of Primer Appearance Defect Based On Machine Vision

The primer is one of the key parts of the bullet,which plays a role in the firing of the bullet.The firing success rate of the bullet is determined by the quality of the primer.The detection of primer defects can guarantee the quality of primer.At present,the primer defect is manually detected,which has problems such as low efficiency,easy fatigue and inconsistent standards.In this paper,aiming at the automatic and noncontact detection requirements of the appearance defects of the primer,the detection method based on machine vision is studied.The work done in this thesis is as follows:1.This paper proposes a detection method for the top defect of the primer.In order to extract the top and side regions of interest quickly and accurately,this paper proposes a fast top-side segmentation algorithm for the primer.On this basis,in order to improve the accuracy of the gap detection at the top of the primer,this paper combines the Euclidean distance and the fast ellipse detection method based on the arc adjacency matrix to determine the gap at the top.Experiments show that the real-time and accuracy of the segmentation algorithm in the detection method in this paper is strong,and the detection method can realize the judgment of the gap defect.2.Aiming at the problem of misjudgment and omission of defect type and damage degree in primer defect detection,a multi-fuzzy reasoning cascade method of primer defect classification and damage degree analysis is proposed.First,the run-length method is used to extract the characteristics of the defect area,average gray value,size and number of points.Second,the input and output sets,membership functions and fuzzy rules of three fuzzy inference subsystems are built.Finally,the three fuzzy inference systems are cascaded to realize the classification and damage degree analysis of various defects.3.In order to detect the filling defects of the bottom powder surface,this paper uses the depth map to be converted into a two-dimensional image,and eliminates the complex noise interference by enhancing the contrast between the foreground and the background.Then,the target area is located and the height information of the medicine surface is extracted,which lays the foundation for the detection of filling defects of the medicine surface.4.Combined with the production process,structural features and defect detection requirements of the primer,the image acquisition subsystem and human-computer interaction interface of the detection system in this paper are designed.The vision system designed in this paper cooperates with electrical and mechanical systems to realize the automatic detection process of primer defects.