Robust orthogonal iterative monocular pose estimation algorithm based on point features

Abstract

Monocular pose estimation is a basic element in computer vision, and monocular pose estimation based on point features, as one of the important branches of monocular pose estimation, is widely used in many fields such as robot positioning, virtual reality, and image precision measurement. When solving the monocular pose based on point features, the number of point features and image noise has a great influence on the estimation accuracy. Therefore, this paper proposes a robust orthogonal iterative pose measurement algorithm that introduces an intermediate coordinate system between the camera coordinate system and the world coordinate system to optimize estimation constraints. Then, the method that uses the least square method to solve the error problem is applied to solving the distance between the characteristic point and the camera's optical center. Moreover, the calculation of the initial value of the camera pose is simplified through the intermediate coordinate system. Finally, the camera pose is optimized by orthogonal iteration. Experiments show that, compared with existing algorithms, the algorithm proposed in this paper is more robust to the number of point features and image noise, and the overall solution accuracy is better.