基于建筑角点的机载和车载点云数据配准

Registration of the Airborne and Vehicle Point Cloud Data Based on Building Corner

提出一种基于建筑角点的机载和车载点云数据配准方法。首先采用随机抽样一致性算法(Random Sample Consensus,RANSAC)对建筑面片进行稳健估计,结合建筑轮廓在二维平面上投影的拟合直线,解算出建筑角点的三维坐标。利用提取到的同名角点,采用六参数转换模型计算机载和车载点云数据间的空间转换参数,进而完成机载和车载点云数据的配准。实验表明,该方法能有效地提取建筑角点,实现机载和车载点云数据的精确配准。

We presented a registration method of the airborne and vehicle point cloud data based on building corner in this paper. Firstly, we used random sample consensus algorithm and the fitting straight of the building contours projected on the 2 D plane to estimate the building patches. And then, based on the extracted homologous building corners, we used six-parameter transformation model to calculate space conversion parameters, and realized the registration of the airborne and vehicle point cloud data. The registration results show that the proposed method can extract high accuracy corner points and achieve accurate registration of the airborne and vehicle point cloud data.