牙颌点云数据三维可视化研究

发布时间：2018-04-30 16:19

本文选题：牙颌 + 点云　；参考：《电子科技大学》2012年硕士论文

【摘要】：在医学辅助诊断领域，口腔三维可视化以其结果能真实的反映病人牙颌病变部位的形态、方便医生以更方便、清晰的方式对病人进行诊断的特点，已经成为临床诊断中一个非常重要的诊断方法，并逐渐成为医学计算机应用的重要研究方向和研究热点。本文中主要研究了牙颌点云的三维可视化问题，针对实际应用需求，提出了牙颌点云去噪，精简，三维重建等一系列操作的可行方案。本文分析了牙颌点云数据的特点，根据实际应用的要求，提出了分两步对点云去噪的方法，选用了单元格法去除了奇异噪声并选用双边滤波法对一般噪声进行了平滑，，两次去噪的方法在去除模型噪声的基础上，最大程度上保证了模型的特征。并在点云去噪过后根据点云密度过高的特点，选择了单元格法对点云数据进行了必要的精简，该方法在保证了模型精度情况下使数据量大为减少，为后面进行点云数据高速三维重建奠定了基础。这些方法在实验中被证实是可行、高效并具有较高精度的。本文着重分析牙颌点云的三维重建问题，根据重建时间，重建精度等实际应用要求，选用了滚球算法对牙颌点云数据进行了三维重建。该算法重建精度高，运行速度较快，能满足应用要求。最后对该算法从精度，复杂度，稳定性等方面和其他算法进行了比较，结果表明滚球算法重建精度高，重建速度较快，运行稳定，并且重建出的模型平滑连续，模型的网格规则合理，能完全满足实际研发的需求。在本文的最后，我们在Windows XP系统上依托于VC++平台，采用wxWidgets，Vcglib和Opengl等类库实现了牙颌点云三维可视化系统。在该系统中我们实现了点云的去噪和重建等功能，并实现了多种数据的输入输出。并应用多种特殊格式的数据对该系统进行了测试分析，结果显示该系统在功能要求，精度要求，时间要求上都达到了目标，正确性和稳定性都能得到保证。
[Abstract]:In the field of medically assisted diagnosis, 3D visualization of oral cavity can truly reflect the shape of the patient's dental and maxillary lesions, and facilitate the doctor to diagnose the patient in a more convenient and clear way. It has become a very important diagnostic method in clinical diagnosis, and has gradually become an important research direction and research hotspot in the application of medical computer. In this paper, the problem of 3D visualization of dental point cloud is mainly studied, and a series of feasible schemes such as de-noising, reducing and 3D reconstruction of dental point cloud are put forward according to the practical application requirements. In this paper, the characteristics of dental point cloud data are analyzed. According to the requirements of practical application, a two-step de-noising method of point cloud is proposed. The singularity noise is removed by using the cell lattice method and the general noise is smoothed by using the two-sided filtering method. On the basis of removing model noise, the method of twice de-noising ensures the characteristics of the model to the greatest extent. After the point cloud denoising, according to the characteristics of high point cloud density, the cell lattice method is selected to simplify the point cloud data. This method can greatly reduce the data volume under the condition of ensuring the model precision. It lays a foundation for high speed 3D reconstruction of point cloud data. These methods are proved to be feasible, efficient and accurate in experiments. This paper focuses on the analysis of the 3D reconstruction of dental point cloud. According to the practical application requirements such as reconstruction time and reconstruction precision, the rolling ball algorithm is selected to reconstruct the dental point cloud data. The algorithm has the advantages of high reconstruction precision and fast running speed, which can meet the requirements of application. Finally, the algorithm is compared with other algorithms in terms of accuracy, complexity, stability and so on. The results show that the reconstruction accuracy of the rolling ball algorithm is high, the reconstruction speed is faster, the operation is stable, and the reconstructed model is smooth and continuous. The grid rules of the model are reasonable and can fully meet the needs of the actual research and development. At the end of this paper, we use the class libraries such as wxWidgetsVcglib and Opengl to realize the 3D visualization system of dental point cloud based on VC platform in Windows XP system. In this system, we realize the function of point cloud denoising and reconstruction, and realize the input and output of many kinds of data. The test results show that the system achieves its goal in function, precision and time, and its correctness and stability can be guaranteed.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R318.0;TP391.41

【参考文献】