Research On Measurement Techniques Of Digital Image Correlation Based On Binocular Vision

In recent years,with the rapid development of science and technology,the three-dimensional measurement of objects' profile and deformation has been widely used in the field of basic manufacturing,aerospace,defense and so on.The digital image correlation method(DIC)based on binocular vision has the advantages of full field,non-contact and low demand for experimental environment.As it also combines the theory of computer vision,the 3D profile information and full field deformation of objects can be measured.Therefore,this method is worthy of study.In this paper,the calibration of binocular vision system and image correlation matching algorithms are mainly studied.Then the measurement system is established to do some verification experiments.The main researches are as follows:The basic theories of binocular vision measurement and camera calibration are studied,including camera imaging model,3D reconstruction,calibration method and so on.In camera calibration process,an improved corner detection method based on Harris algorithm and a corner sorting method are proposed.The detection method improves the deficiency that the appropriate threshold in traditional Harris corner detection algorithm needs to be set manually,and the sorting method can sort all the corners automatically and accurately.After studying and comparing the calculation accuracy and speed of several integer-pixel and sub-pixel correlation matching algorithms,the fast ZNCC algorithm and the inverse compositional algorithm are chosen as appropriate ones.In addition,in stereo matching,when iteration is not convergent due to the initial value,by combining traditional initial value calculation with the displacement gradient between two points,the number of error matching points can be reduced.In sequential matching,by setting a initial calculation point,the computational efficiency of the whole field can be improved,which further reduces the the calculation time.The digital image correlation measurement system based on binocular vision is established and the relevant algorithms are finished.Then experiments of 3D complex profile reconstruction,semi cylindrical radius fitting,rigid body displacement and sample stretching are conducted.The effectiveness of the measurement system and algorithms are verified.