Research On Visual Imaging And Defect Detection Technology For Aerospace Seals

O-ring is a flexible part,the force will be deformed.In the use of spacecraft,aircraft and other high reliability requirements,it is necessary to strictly measure and detect its size and surface defects.At present.The common practice in the industry is to measure the size of the manual operating tool microscope and the human eye to use a magnifying glass for defect detection.This method has low efficiency and unstable results.The purpose of this paper is to solve these problems by studying the vision imaging and defect detection technology of space sealing ring.Using machine vision technology to detect the surface defects of sealing rings used in aerospace systems,due to the existence of highlight region on the surface of the sealing ring,the pixel value of the defect tends to saturation,which makes the defect difficult to be detected.In order to avoid the adverse effect of the highlight area on the sealing ring surface on defect detection,according to the principle of surface restoration,the height information of the sealing ring is calculated by collecting many two-dimensional images,and the 3D reconstruction of the sealing ring surface is realized.Finally,the defect is detected by determining the height of pixels.This paper first discusses the method of 3D reconstruction based on the theory of photometric stereo vision,and then points out that the height calculated by this method is greater than the true value at the highlight area of the sealing O-ring.Combined with the principle of surface restoration,the normal vector at the highlight region of the sealing ring calculated by photometric stereoscopic vision does not match the true value.In order to obtain a more accurate surface normal vector,using the improved voting algorithm to calculate the normal vector.The improved voting algorithm is based on:when the light source rotates,the diffuse reflection intensity of a point of reflection on the surface of the object and the angle of the light source obey the sine curve.Based on the principle of Hough transform,the sinusoid corresponding to each pixel is determined.The surface normal vectors corresponding to each pixel are calculated through the sinusoidal parameters.Then a bright material ball is drawn to verify the value of the normal vector calculated by the improved algorithm.Then,the simulation study of two groups of seal surface with the highlighted area and without the highlighted area,analyses the difference between the photometric stereo and improved voting algorithm in the vector solution,the height change before and after mutation vector comparison highlight regions,the section height curve of the sealing ring recovered by the two methods was drawn and compared respectively.Finally,an optical darkroom image acquisition system is set up to capture the sequence images of the actual sealing O-ring.The surface normal vector is calculated by the improved voting algorithm,and the sealing ring is reconstructed in three dimensions,the defect detection is carried out by using the surface height information.Through the research of this paper,we can avoid the negative influence of the surface highlight area of the space sealing ring on the defect detection algorithm,and realize the 3D detection of the defect.