The Study Of Digital Image And Its Application In Deformation Measurement

Digital image correlation method is a non-contact full field optical measurement method developed in recent twenty years which has some unique advantages compared with traditional optical measurement and other measurement methods.It combines modern information technology,computer image processing technology and computer vision technology.The research of two-dimensional digital image has already matured nowdays,its development is mainly directed toward the research of high-efficiency and high-precision and engineering application,this method is limited to the measurement of in-plane deformation of objects.The research of 3D deformation measurement technology has recently attracted much attention.The image correlation method is based on the cross-correaltion between the artificial or natural speckle fields distributed randomly before and after the deformation of object surface to establish the deformation information of the object.Aiming at the limitation of two-dimensional digital image correlation method application,it is necessary to introduce stereo vision measurement technology into two-dimensional image correlation method and study the three-dimensional shape and deformation measurement of objects.This paper focuses on the two-dimensional image correlation search algoritm and the key technologies of 3D image correlation method and the development of related measurement software in the system.We study the key research algorithm in research of 2D image correlation method and the correlation coefficient fomula is determined.Based on the existing algorithm we propose an improved correlation search method that can be used for the search of surface deformation information and carry out related mesurement experiment,the results show that the proposed method can satisfy the mesurement efficiency and accuracy requirements.On the basis of determining the related methods of two-dimensional search,research on the key technology of 3D image correlation method is carried out.We study the camera calibration method which is suitable for the research need,with the help of vision processing library,the camera parameters are determined,meanwhile,with the introduction of camera lens shear parameter,the camera intrinsic parameter nonlinear optimization is carried out so as to improve the calibration accuracy.The key matching technology of stereo vision measurement is studied,the images are processed by wavelet denoising preprocessing,and the outer epipolar line constraint of camera model is used for stereo matching.After the initial matching is completed,the RANSAC algorithm is introduced to correct the mismatch and complete the stereo matching of binocular images.On the basis of the existing theoretical research,a friendly,simple and easy-to-use correlation measurement software is developed.The relevant deformation tests are carried out to verify the feasibility.Finally,related error analysis is carried out and a feasible scheme to reduce the error is put forward.