Research on Fusion and Registration Method for High-Resolution Satellite Image and Vehicle Lidar Data.

In this paper, to solve the problem of lacking of road information caused by ground object occlusion, a registration and fusion method of high-resolution satellite image and vehicle point cloud data is proposed. Firstly, the road surface and crash barrier are extracted by using the filtering algorithm of joint gradient and elevation. Secondly, the Canny algorithm is used to extract road boundary on satellite images, and the plane is selected to extract linear points according to the elevation features of crash barrier and road boundary. Thirdly, the nearest neighbor point cloud iteration method is used to realize the matching of linear points with the same name. Finally, the high-resolution image and DEM are combined to generate a 3D model, and the vehicle point cloud data is registered with the three-dimensional model according to a rotation matrix, so as to improve the efficiency of high-precision map construction. The experimental results show that this method can effectively realize the registration of high resolution image and vehicle lidar data, combine the complementary advantages of high resolution image and point cloud data, improving the integrity of vehicle point cloud data, and alleviate the problem of high-precision map construction caused by occlusion to a certain extent.