Real time pose estimation based on extended Kalman filter for binocular camera
Wenhai Liu, Jie Hu, Yi Fang, Quanquan Shao, Kai Zheng, Guo‐Niu Zhu · 2016
Real-time pose estimation is a challenge in multi-camera vision system due to the demand of rapid response, high accuracy and robustness. Although some works based on multi-camera have been proposed, few works have regarded multi-camera as a fixed integration, which is easier to apply in real robotic application than eye-in-hand/eye-to-hand configuration. This paper proposes a novel Kalman-filter-based pose estimation algorithm for binocular camera that is regarded as a fixed integration. Centralized fusion method is employed to conduct the fusion of the eye-to-hand binocular camera. Extensive simulations are conducted to validate the effective of the proposed fusion method. The results show a higher accuracy and precision compared with traditional algorithms.