Pose Tracking of 3D Target Based on Iterative Weighted Least Squares

Abstract

In order to achieve precise tracking of 3D target pose, a new pose-tracking method based on iterative weighted least squares is proposed. Aiming at the pose tracking problem of the weak texture artificial target, the line is chosen as the tracking characteristic. According to the distance between the image line and the projection line of the model, the pose error function of the end point and the midpoint of the line is considered synthetically. For the influence of noise on image line extraction and the mismatch between corresponding lines, we obtain the robust weight by the angle between the error function and the corresponding line. Finally, the problem of solving pose parameters is transformed into a weighted least squares problem, and iterations are added to ensure the accuracy of the solution. Simulation image experiments show that the proposed algorithm has high accuracy and good stability. The real image experiment shows that the algorithm can overcome the interference of complicated environment and realize the exact solution of 3D target pose.