中文摘要：

激光扫描仪（Laser Scanner，LS）获得的点云数据由海量全离散的矢量坐标点构成。由于点云数据没有明显的形体信息和拓扑关系信息，考虑数据本身蕴含的丰富信息，采用改进后的自适应紧支撑径向基函数（Adaptive CS-RBF）算法对点云数据进行曲面重建。在介绍紧支撑径向基函数数学模型的基础上，采用激光扫描获得同济大学孔子头像点云数据进行实例建模，通过比较不同建模参数获得的模型，分析该方法的应用效果。

英文摘要：

The sea volume of point cloud can be obtained by Laser Scanner, which is consisted of scattered vector point. However, the point cloud is lack of distinct geometry and topological information. In order to build a real three dimension surface model by using point cloud, the method of compactly supported radial basis function is discussed by considering internal information among point cloud in this paper. The principle of compactly supported radial basis function is given firstly as the mathematic basis. Then, the flow chat of model application is explained before the case studying. Point cloud of Kongzi portrait in Tongji University is used as the case study to build the three dimension surface model. Comparison is conducted among the established model with different model parameters,and some result is obtained with a detail analysis.