Image moment-based planar object pose estimation

To address the error problem caused by the coupling of in-plane object pose estimation, this paper proposes an image moment-based planar object pose estimation, which simplifies the computational process by adding a virtual rotation process of the object plane and achieves a closed solution by completely decoupling the rotational and translational components of the interaction matrix. On this basis, a control scheme based on a monocular camera is designed to verify the results of the in-plane object's positional estimation. The experimental results show that this method can realize the exponential decrease of the error of the plane parameters and the velocity of the object, and the velocity of the object is stable and has good robustness.