Abstract: At present, the problem of real scene 3D modeling lies in the deformation of some ground object models and the fuzzy texture at the bottom of buildings caused by the missing image. In view of the shortcomings and defects of oblique photography and vehicle-borne LiDAR point cloud data sources in the construction of high-quality and high-precision urban 3D model, this paper explores the main technical process and method of using TerraSolid software to merge the point clouds obtained from the two and then model. The results show that the laser point cloud makes up for the geometrical structure of the oblique photogrammetry blind area, making the ground of the 3D model more smooth, and the edges and corners of the bottom of the building more sharp and clear. The organic combination of vehicle-borne LiDAR point cloud and oblique photography with complementary advantages can effectively make up for the defects of using a single data source to make a 3D model and improve the precision of the model.

Key words: vehicle-borne LiDAR, oblique photographic image, fusion modeling, TerraSolid, 3D real scene