Aiming at the problems of low positioning precision and discontinuous positioning caused by the error accumulation generated by the visual localization process and the non-line-of-sight(NLOS)error generated by the ultra-wideband(UWB)and localization influenced by occluded objects in indoor environments

localization method a combined binocular vision and UWB is proposed

which adopts a factor graph data fusion algorithm to fuse the ranging data of UWB and the position information of visual localization to achieve high-precision continuous localization in indoor environment.The experimental results show that the combined localization method can effectively reduce the cumulative error of visual localization and the NLOS error of UWB

and the maximum localization error is reduced by 47.6% and the localization precision is significantly improved compared with the extended Kalman filtering(EKF) combination method.