地形图辅助的建筑物地面和机载点云特征配准

Feature based Registration Between Terrestrial and Airborne Point Cloud Assisted by Topographic Maps

摘要:

机载和地面激光扫描数据配准的实质是坐标转换问题.以建筑物的立面为研究对象，基于地形图的建筑物线性特征，提出建筑物线面特征约束的地面和机载点云数据配准方法，实现机载和地面点云数据的配准及其向地形图坐标系的转换.采用拟合方法，获取地面与机载点云、地形图的建筑物线性特征参数.依据特征之间的空间关系，建立特征参数与坐标转换模型的关系，分别实现地面与机载点云向地形图坐标系的水平转换.垂直转换则通过建筑物水平屋顶边缘的高程相对配准以及控制点处的高程绝对配准计算.以上海海洋水族馆点云为例进行实验验证，结果表明该方法可以实现地面和机载点云数据配准.

Abstract:

Coordinate transformation is the essence of registration between airborne and terrestrial LiDAR system. The traditional coordinate transformation needs to obtain several common points and calculate the model parameters by those points. However, common points are hard to select in terrestrial and airborne point cloud due to different platforms and discrete sample mode. Taking building profile features as objects, a features constraint registration method was proposed in this paper to achieve the registration between point clouds and topographic maps. First, the feature line and feature plane of building facades were obtained from airborne and terrestrial point cloud separately. After being projected to the horizontal, the plane transformational parameters were calculated according to the relationship between model parameters and feature parameters. Next, the vertical parameter were computed by slicing the top layer of point cloud and elevation of the control points. Finally, the method was verified by a case study in Shanghai Ocean Aquarium.