Conjugated-Point Positioning Method Based on Multiview Images

The performance of the direct georeferencing method depends on the accuracy of the position and orientation data. The accuracy of the attitude data has a particularly significant impact on the positioning results. This paper proposes an object point positioning method that does not require attitude information. In the proposed algorithm, only the position information of the perspective center is necessary. The basic principle is that the error equation of each image can be established according to two visible object points and the image perspective center. The angle between the rays connecting these points and the center is the observation value that can be calculated from the image points and camera parameters. In contrast to the spatial intersection method, this method does not need the attitude information to solve for the coordinates of object points. This paper analyzes the main factors affecting the accuracy through simulation experiments. Using a DJI Phantom 4 RTK unmanned aerial vehicle in experiment, the accuracy of positioning four points can achieve 0.2 m when the flight altitude is 60 m and the flight radius is 40 m. On the experiment of positioning N object points, the positioning accuracy can achieve 0.105 m. The method in this paper can achieve the same positioning accuracy as the space intersection.