Key Technologies Research And Application Of Micro Digital Dental3D Scanning System

The application of digital technology in the medical field of dental restoration has led to atechnical revolution of dental restoration. The replacement of the cumbersome, time-consuming, andrough manual methods by simple, efficient, and accurate digital process, not only improves theaccuracy of the restoration, but also greatly reduces the suffering of patients. Being the basis of dentalrestoration, digital measurement directly determines the quality of restoration. Combined with thedemands of dental clinic, this dissertation thoroughly studied the digital optical dental measurementmethod based on machine vision including imaging processing, camera calibration, phase computing,reconstruction, registration, software&hardware implementation, etc. Two dental measurementprototype systems: D3DscannerI and D3DscannerII are also developed. The dissertation’s maincontributions are as follows:1. The dental measurement model is constructed and is analyzed from the perspective of systemcalibration and reconstruction respectively. The overall framework and routing of the development ofdental measurement system is devised.2. System calibration solution based on the overall measurement error is developed. First, thecalibration pre-treatment technology based on CCD noises and image definition is used to ensure theaccuracy of the input data. Then the nonlinear parameters of the system is calibrated with leastsquares and the distribute convergence algorithm based on the RAC. Thirdly, considering thegeometry restrain of the system structure, the overall optimization calibration method is established.Finally, the error between the ideal system model and the actual physical model is taken as the targetoptimization function to optimize the system again. The developed three-step calibration method canobtain calibration system parameters with high accuracy, and the third step of the optimization meetsthe special requirement of the calibration of the intraoral micro-measurement system.3. A new robust phase unwrapped algorithm for discontinuous teeth shape measurement isdeveloped. First, this paper elaborates and extends the phase quality factor from physical model, anddevelops a new modulation-Laplace quality factor based on mathematical statistics and imageprocessing technology. The phase unwrapping are achieved along the absolute phase quality map,realizing the settlement of the correct phase of complex discontinuous teeth morphology.4. The phase shifting algorithm immune to nonsinusoidal phase error and phase-shifting methodwith the absolute phase marker embedded in phase stripes are proposed. The former reduces thenonsinusodial phase error to1/9of the traditional compensation algorithm by compositively utilizing the extremal convergence properties of trigonometric and the zero point of phase computing. Thealgorithm greatly improves the system measurement accuracy. The latter embeds the new designedabsolute phase marker into one of the phase fringe images, and reduces the phase shifting fringesimage from5to4, which avoids the need of additional absolute phase marker map. The methodimproves at least20%efficiency compared with the traditional methods.5. The registration solution based on double shaft and ICP is developed for small oral dentalmeasurement system. First, the eigenvalue method is used to achieve the initial parameters of rotationand swing axis, then the iterative convergence algorithm based on the integration of point-to-point andpoint-to-surface distance metric function is used for accurate registration.6. The optimization registration algorithm is proposed for miniature intraoral dental measurementsystem. Based on the homography between a multi-source image and3D measurement points, thematching relationships of two-dimensional (2D) texture gray images andtwo-and-a-half-dimensional(2.5D) range image are constructed using the scale-invariant featuretransform algorithms firstly. According to the index of overlapping image and the local3D bordersearch method, multi-view registration data are rapidly and accurately integrated. At last, theimproved ICP algorithm is used to acquire accurate and stable registration.7. Two prototype dental measurement systems, namely D3DScannerI and D3DscannerII basedon machine vision is designed and realized with the key algorithms and techniques described above.The system model including accurate system calibration, rapid measurement, multi-view registration,3D reconstruction, cloud polygonization and display algorithm et al. have all been coded,implemented and integrated with VC++for the primary functionality of the system. It has beendemonstrated that the complex teeth shape surface point cloud can be obtained fairly well by theprototype system. Two prototype dental measurement systems developed can meet the requirement ofdental clinic basically.