飞机装配的人工环节中可利用 AR 技术的虚实融合功能将装配指令叠加显示在装配场景中。高质量的虚实融合效果需要 AR 系统具备精确的空间定位能力。本文提出一种基于平面标志物信息约束的同时定位与建图(SLAM)空间定位方法。AR 系统首先通过 SLAM 方法建立初始 AR 地图;随后通过识别 AR 地图关键帧中平面标志物并利用标志物信息约束对关键帧和地图点的变换及优化建立高质量 AR 地图;最后,AR 系统通过跟踪新的 AR 地图进行空间注册和视点跟踪,从而实现装配场景坐标系下的高精度空间定位。验证试验证明该方法有效提升了 AR 系统空间定位的精度。
Abstract:In the manual link of aircraft assembly, the combination of virtuality and reality by the AR technology can be used to display the instructions in the scene of assembly. High quality combination of virtuality and reality requires the ability of accurate localization of the AR system. This paper proposes a method of localization by using simultaneous localization and mapping (SLAM) based on the constraint of planar markers for AR-based assembly. Firstly, the AR system generates the original AR map using the SLAM method. Then, it detects the planar markers in the key frames of the AR map, and the constraint of the markers is employed on the transformation and the optimization of the key frames and the map points to build the high-quality AR map. Finally, the AR system tracks the new AR map for registration and view point tracking to enable high accuracy localization under the coordinate system of the assembly scene. The verification experiments demonstrate that the method can effectively improve the accuracy of localization of the AR system.
霍志宇. 高精度增强现实辅助装配空间定位技术研究与应用[J]. 航空制造技术, 2022, 65(18): 105-110.
HUO Zhiyu. Research and Application of High Accuracy Localization for Assembly Assisted by Augmented Reality[J]. Aeronautical Manufacturing Technology, 2022, 65(18): 105-110.