Accurate 6-DoF pose sensing and dynamic tracking enabled by a metalens-integrated camera

Accurate six-degree-of-freedom (6-DoF) pose sensing is essential for autonomous systems, intelligent robotics, and the Internet of Things. Conventional pose-sensing approaches often rely on bulky imaging modules and complex visual pipelines, limiting their deployment on compact and embedded platforms. Metalens-integrated cameras provide an ultrathin optical front end, but their reduced contrast and imaging blur make pose estimation unstable. Here, we demonstrate 6-DoF pose sensing and dynamic tracking using a metalens-integrated camera assisted by an ArUco marker tailored to its imaging characteristics. Robust four-corner correspondences are directly extracted from metalens images, and a calibrated planar PnP solver enables accurate pose estimation. Static measurements yield mean absolute translation errors of 1.43, 1.08, and 4.71 mm along the x, y, and z axes, respectively, with mean yaw, pitch, and roll errors of 1.41°, 6.37°, and 9.06°. Dynamic experiments further show stable continuous marker tracking, highlighting the potential of this compact localization framework for embedded pose-aware sensing.