Research On Autonomous Localization Technology For Immersive Measurement

In the manufacturing process of large equipment such as ships and spacecraft,it is necessary to complete the assembly and docking of pipelines inside the cabin.Existing large-scale measurement instruments such as indoor GPS and laser tracker have poor measurement flexibility and weak man-machine interaction in complex and restricted internal environment.In this paper,an immersive measurement method is proposed to make measurement personnel quickly integrate into the field measurement environment,in which autonomous localization technology plays a key role in the immersive measurement.Aiming at the problems of low accuracy and poor adaptability of autonomous localization in complex measurement environment,a fast localization method based on multi-camera system combining artificial beacons and visual SLAM(Simultaneous Localization and Mapping)is proposed.Beacon localization algorithm and visual SLAM algorithm are designed respectively to realize the whole space localization in different visual situations.Artificial beacons need not cover the whole space layout and don’t need to have ID information.Each camera runs different localization algorithm.A data fusion algorithm is designed to realize the interaction between the pose data of the two localization algorithms,so as to restore the scale factor of monocular SLAM and establish the transformation relationship between the two algorithms.The multi-camera localization system is installed on the wearable helmet,which enables measurement personnel to carry out high-precision positioning and achieve immersive measurement.The overall design of helmet localization system,the principle and implementation of localization algorithm were analyzed and demonstrated.Finally,an experimental platform was built to verify the feasibility and effectiveness of the proposed method.The experimental results show that the positioning accuracy and environmental adaptability can meet the requirements of the immersive measurement system in complex and restricted environment.The main research work of this paper is as follows:(1)Aiming at the problems of poor measurement flexibility and man-machine interaction of existing measurement instruments in the field of assembly and inspection of large equipment,an immersive helmet localization system was designed,and the overall composition and algorithm design of the helmet localization system were introduced.(2)Beacon localization algorithm is designed,the pose representation method using Lie algebra is introduced,the initial value of pose is obtained based on constant velocity model,the principle of feature point matching based on fast nearest neighbor is described,and the method of solving the exact pose of unconstrained optimization problem based on bundle adjustment is designed.The visual SLAM algorithm is designed,the feature point extraction strategy based on quadtree is proposed,the pose solution method of tracking the adjacent frame and local map is described,and the closed-loop detection strategy is discussed based on similarity comparison of Bag-of-Words vector.(3)Aiming at the problem of the inconsistency of coordinate system between two localization algorithms and the scale uncertainty of monocular SLAM,a data fusion algorithm is studied.When the visual condition of the system changes,beacon localization algorithm is restored using the pose result of SLAM,and the global optimization of SLAM algorithm is carried out with the latter pose.(4)The software and hardware experimental platform was built.VSTARS is used to construct a 3D control field.The positioning accuracy and robustness of the proposed algorithm was verified by experiments using high-precision translation stage and T-mac of laser tracker as measuring datum.