The Evaluation Of Speckle Quality And The Research Of Stereo Matching In Three-Dimensional Digital Image Correlation Method

Nowadays,in medical,mechanical,aviation and other fields,a large number of workpieces with complex materials will be used.After long-term operation,these workpieces will produce a certain degree of wear and tear.Therefore,we need to use a convenient,full-field and strong adaptability method to measure the deformation information of materials regularly.With the continuous development of measurement technology,optical measurement has been widely used in various fields.Digital image correlation method is an optical measurement method based on image correlation calculation of object surface.It has the advantages of non-contact,full-field measurement and strong environmental adaptability.On the basis of the two-dimensional digital image correlation method,the three-dimensional digital image correlation method can accurately measure the three-dimensional information of the object,reconstruct the three-dimensional shape of the surface of the object,and realize the in-plane and outof-plane displacement and deformation measurement of the object through the dualcamera measurement system,combined with image processing and computer vision technology.This paper mainly studies three aspects of three-dimensional digital image correlation method: quality evaluation of speckle image,camera calibration,stereo matching and three-dimensional reconstruction.The main research work is as follows:In order to ensure the accuracy of speckle image in stereo matching and threedimensional reconstruction,the quality evaluation of speckle image is studied in this paper.Based on the neural network,a new method for speckle quality evaluation is proposed,which takes the average gray gradient,the energy,contrast,correlation and entropy of gray level co-occurrence matrix as evaluation parameters.The validity and accuracy of this parameter are verified by experiments.In camera calibration,according to the camera calibration model,an improved Harris corner detection algorithm is proposed,which realizes the automatic corner detection of calibration plate.On this basis,the accurate sub-pixel coordinates of corner points can be obtained by modifying the whole pixel of corner points,and then the internal and external parameters of left and right cameras can be calculated to establish the basis for stereo matching and three-dimensional reconstruction of images.On the basis of speckle quality evaluation and camera calibration,image stereo matching and three-dimensional reconstruction are studied.An improved stereo matching algorithm combining RANSAC algorithm and SURF algorithm is proposed,which can quickly and accurately screen out the feature points of the image and generate matching pairs of feature points.Then the three-dimensional reconstruction of the object is realized by the camera parameters inside and outside,and then the displacement and deformation of the object can be obtained,which provides measurement information for engineering application.