基于激光扫描仪的点云模型的自动拼接

Automatic Registration of Point Clouds Based on Laser Scanner

利用非接触式三维激光扫描仪对目标进行旋转扫描,获得目标的不同旋转位置的空间数据和数字影像,对相邻的两幅立体影像利用匹配方法寻找同名点,并映射到空间方,同时利用相对定向和空间相似变换分别剔除像方和空间方的粗差,以获得相邻两个坐标系准确的空间相似变换参数,根据空间相似变换参数将位于相邻的两个坐标系的空间点统一到同一个坐标系中去,完成三维数据的配准。

This paper uses the non-contact 3D laser scanner to acquire the 3D point clouds and digital images, then searches the correspondent points in the stereo images through image matching. Then the correspondent points are mapped to the space points, so as to get the correct spatial similarity transformation parameters, and unify the points siruated at two coordinates to the same coordinate system. In this way, the 3D registration can be completed.