3D Digital Measurement Of Objects Based On Binocular Vision Systems

With the development of modern industry, the appearances of diversified products and improvement of people's living standard, the traditional measurement content and method can't already meet the demand, the outline and outline data of object is becoming a new and important research direction in the measurement field. Especially the 3D measurement of irregular freedom curved surface and shape, for instance, appearance of the mould, turbine design, engine outer cover, automobile appearance, etc. How fast and high accurately to obtain the object 3D surface data becomes a difficult problem to be urgently solved in modern industrial development.The dissertation introduces the demands and application domains of three-dimensional contour measurement technology, and gives analysis and comparison of contact and non-contact 3D measurement technologies. It had many advantages for binocular vision systems such as high speed, high precision, non-contact, non- destruction, simple operation and low cost. To meet the need of application, Zhang's flexible new method to calibrate the camera is improved, and a new and accurate method for camera calibration plane's corners detection is proposed, and uniform calibration mathematics model is established; various feature points detectors are compared and eventually the 3D digital measurement of objects is realized with the effect detector and stereo matching technology based on epipolar constraint. The dissertation deals with some problems that are theory as well as methods, project application as well as experiments, and it involves such multi-disciplinary fields as coordinates measurement technology, computer technology, image processing technology and matrix theory.etc. Both theoretical and practical researches and discussions were taken, and some parts of the experimental validation were finished.