Abstract: Based on the point cloud data registration method of feature primitives, this paper uses the control points to roughly match the airborne and vehicle point cloud data, and constructs the feature point matching algorithm model which takes the gradient and color feature and the feature group pair into account. Then, it realizes the accurate registration of airborne and vehicle point cloud data with translation and rotation transformation parameters obtained by fitting the plane feature. On this basis, the technical process of matching the top and facade features of the multi-angle point cloud data fusion, the point cloud data and the optical image texture information are established, and the building (structure) 3D fine modeling is realized, finally the validity of the proposed method is verified by an example.

Key words: multi-look laser point cloud, matching, fusion, 3D modeling