Study Of Vision Deformation Measurement Technologies Based On Digital Speckle Correlation Method

With the rapid development of manufacturing, the measurement research of 3D deformation and surface profile for various loaded materials and structure plays a significant role in the fields of material test, intensity experiment, quality control, reverse engineering, and so on. Vision measurement gets more recognition and investigation for its quality of non-contact, high sensitivity, high measurement accuracy and high automation in recent years. Digital speckle correlation method (DSCM) is a kind of optical measurement technique with all field of view, non-contact, and an image measurement technique by applying computer vision, which has been used in experimental mechanics, various engineerings and other research fields. But the technique can only be used for the in-plane displacement of a plane object, it is most essential to continue 2D digital speckle correlation method (2D-DSCM) to 3D digital speckle correlation method (3D-DSCM) in order to get 3D deformation measurement results for plane or curve structures.The subject"Study of vision deformation measurement technologies based on digital speckle correlation method"is to study a new 3D deformation measurement approach with all field of view and non-contact based on the present researches of 2D-DSCM, and introduce binocular stereo vision technique. For the study of deformation measurement, we must solve some key technologies, such as search accuracy and velocity of 2D-DSCM, matching for 3D-DSCM, and precision calibration for the camera parameters, etc.In the study of 2D-DSCM, some key techniques have been researched. Correlation coefficient with acuti-peak value and handy computation is confirmed in order to improve search velocity and computation precision. Adaptive genetic algorithm for correlation search to get in-plane deformation values is presented based on analyzing present correlation computation methods, which is more accurate and quicker than traditional search methods. Sub-pixel evaluation method is confirmed. Adaptive GA search method is validated by the simulated speckle images, the results indicate that the method is better than other search methods with rapid search and high precision. In the study of 3D-DSCM, stereo vision measurement technology is introduced in the digital speckle correlation method. The solving method for 3D coordinates of space points is discussed. In order to solve the main problem of stereo matching in the stereo vision measurement technology, this paper presents a quick and exact stereo matching method for the digital speckle correlation method, which is realized by combining the principle of epipolar geometry and grey correlation for the same meeting points matching. The values of the 3D deformation can be acquired according to the same point with 3D coordinates before and after deformation when the stereo matching is finished.Camera calibration is the most important step for ensuring measurement accuracy, this paper presents a valid camera calibration method based on 3D standard space points and improved two-step methods. The 3D standard coordinates of calibration space points are carried out by an infrared LED as optical feature point, which is moved with the coordinate measuring machine (CMM). The method involves initial values for most calibration parameters with the points near the light axis because of their little distortion by the linear estimation, followed by some iterative solutions for the other parameters with all space points by minimizing the distances from all the 3D space points to the corresponding optical rays generated by their projection image points and the center of camera. The proposed technique overcomes the shortcomings of the machining and calibration of large stereo calibration pattern and lower precision of plane pattern. By optimizing the distances from all the 3D space points to the corresponding optical rays, the method assures the spatial distance error is minimal, and enhances the calibration precision. Because of consisting of linear optimization parameter estimation and nonlinear refinement, the calibration process can quickly and availably converge to the totally optimization.The general experiment of the measurement system for actual measured object approves the theory is valid, the algorithm is feasible, the system is of high precision and can satisfy the demand of the deformation measurement, and has good application.