Abstract:

Based on the discrete 3D data points on the object surface, a human-computer interactive object simulation algorithm is proposed. Firstly, the mean and variance of probability statistics is adopted to get rid of the noise. Then, Gaussian projection is applied to construct initial grid of the data points by using the concept of longitude and latitude coordinates of earth geography. Finally, Bezier surface is used to find the new control points and subdivide initial network. In the whole simulation process,  keyboard control functions of OpenGL is adopted to realize human-machine interactive modeling, which sets the lighting and material on the modeled object. Experimental results show that the proposed simulation modeling method is effective, which can be directly applied to the repair of ancient buildings, 3D printing and other fields.

Key words: removing noise of point cloud, Gaussian projection, grid optimization, Bezier surface, human-computer interaction

摘要：

基于物体表面的离散三维数据点，提出了一种人机交互式物体仿真建模算法。使用概率统计中均值与方差的知识对数据点进行去噪处理，借助地球地理经度、纬度坐标的概念采用高斯投影的方法对数据点进行初始网格的构建，借助Bezier曲面寻找新的控制点来对初始网格进行插值细分。在物体的整个仿真过程中使用OpenGL中的键盘控制函数实现人机交互式的建模，并且对建模后的物体进行光照、材质等的设置。实验结果表明提出的仿真建模算法是真实有效的，可以直接应用于古建筑物的修复、3D打印等领域中。

关键词: 点云去噪, 高斯投影, 网格优化, Bezier曲面, 人机交互