Six-DOF Pose Estimation Based on Multiple CArUco Markers

Accurate and real-time 6-DOF (degree of freedom) pose estimation is crucial for robot motion and operation. Currently, marker-based pose estimation relies heavily on image processing and is widely used. However, environmental factors such as lighting changes, occlusions, and image blur can degrade accuracy and robustness. To address these challenges, this paper proposes a novel 6-DOF pose estimation method using multiple CArUco (Circle_ArUco) markers. First, a new marker design adds four circular features to traditional ArUco markers, improving corner point correction and pose accuracy. Second, to reduce the impact of outliers in PnP-based pose estimation, a RSQPnP (RANSAC_SQPnP) algorithm is introduced, combining RANSAC with Gauss-Newton optimization for enhanced accuracy. Finally, a custom octahedral fiducial object enables multi-angle detection by a single camera, improving robustness under occlusion. Experiments confirm that the proposed method significantly enhances pose estimation performance.