3D map generation methodologies have been studied for many years. They are usually based on images acquired by aerial or land vehicles. These methods essentially depend on location information where images are captured. Generally, conventional photogrammetric operation methods rely on Ground Control Points (GCPs) for reliable georeferencing information, which is also called Indirect Geo-referencing (IG). However, it requires excessive cost and time to survey GCPs beforehand. On the other hand, the Direct-Georeferencing (DG) methods can reduce the timeconsuming procedures considerably with high-quality Global Navigation Satellite System (GNSS) and Inertial Navigation System (INS). However, it is difficult to meet the accuracy requirement for many mapping applications with low-cost Micro Electro Mechanical Systems (MEMS) Inertial Measurement Unit (IMU) and GNSS. In this paper, we propose an efficient method to compensate navigation solutions and 3D information utilizing monocular vision, GNSS and INS. By the proposed method, the navigation solutions are compensated by the feature points between two consecutive images and the absolute 3D feature coordinates of point clouds are obtained at the same time. An experiment was carried out to evaluate the feasibility of the proposed method in a dense urban area. The experiment results confirm that the proposed method can compensate the key navigation variables and 3D information.

Keywords: GNSS, INS, VISION, feature point, point cloud