| Eye is one of the most important organs of sense. Research shows that 80% of the information gained outside is from our eyes. Healthy eyes play a very great role in the quality of our lives.The structure of eye anterior segment is directly related to many diseases, and these diseases can directly display on the parametric variation of eye anterior segment. Therefore, these parameters are important on diagnosing, treating method, the observation on disease development and the evaluation of treating in clinical practice. Consequently, we plan for development a 3D analytic system about eye anterior segment based on Scheimpflug imaging theory.To achieve our research goal, the calibration methods for system were researched which lay the foundation for next research. According to principle of digital close-range photogrammetry (DCRP), homologous points (HPs) are the basic data for system. HPs are originated from 3D object point and 2D image point.The main issues of calibration were researched in this dissertation. Some methods were improved and some are novelly proposed, which are including:1: The calibration of camera is fully based on DCRP method. The geomtric parameters are firstly estimated using Direct Linear Transform method. Then the bundle adjustment is applied to optimize parameters, for better accuracy. The projector is calibrated in the same way as a virtual camera by capturing the calibration points in the projecting plane using two-dimensional phase demodulation.2: A technique of auto-detecting feature object was implemented, which is used in label-orientation method. The technique can supply high-precision sparse HPs for geomtric calibration. Otsu binary method was used to separate elliptical block, labeling connected components with Union-Find structure.3: The 3D image was displayed in the visual interface according to calibrated parameters and object's coordinates. The precision of result was analysed, which shows the efficiency and stability of proposed methods. |
PDF Full Text Request