Accurate Pose Estimation Method using ToF-Stereo Camera Data Fusion

Recent studies on sensor pose estimation have expanded due to their applications across a range of fields, including augmented reality (AR), virtual reality (VR), and robotics. Within this context, we present an innovative approach that diverges from traditional techniques involving images or IMU sensors. Instead, we used pre-reconstructed 3D maps to enhance the precision of pose estimation. By training on these 3D maps, we can track objects using data from ToF-Stereo fusion and utilize this information to calculate precise positional coordinates within indoor environments. Through experimentation with a ToF-Stereo prototype, we collected data and tested the algorithm to validate the proposed method.