面向大型客机翼身对接的增强现实跟踪注册方法研究

Research on Augmented Reality Tracking Registration Method for Wing-fuselage Docking of Large Passenger Aircraft

中文关键词: 翼身对接  增强现实  跟踪注册  模板匹配  果蝇优化算法

英文关键词: wing-fuselage docking  augmented reality  tracking registration  template matching  fruit fly optimization algorithm

中文摘要:

      飞机大部件的装配质量往往对整机的装配质量起决定性作用。随着增强现实技术的发展，可对大部件装配对接过程涉及的多源异构数据以图形或数字形式进行直观展示，辅助工艺人员更好地完成复杂装配任务。为了将虚拟信息叠加到现实世界中，需要对大部件的位姿进行准确定位。然而在实际翼身对接过程中，大部件本身的复杂性以及装配环境的限制导致传统的跟踪注册方法在应对这些挑战时出现定位精度不高、计算效率较低的问题。因此，文中提出了新的跟踪注册方法。首先，基于Unity Perception构建了图像数据集；然后，基于Harris和尺度不变特征变换法（Scale Invariant Feature Transform, SIFT）实现了特征点识别和模板匹配；最后，基于果蝇优化算法优化了匹配阈值，并通过翼身对接模拟实验平台验证了该方法的有效性，新方法的总误差百分比平均值仅为2.006 9%。该方法解决了大部件跟踪注册定位中的关键问题，为增强现实技术在飞机装配领域的进一步应用提供了更为可靠的支持。

英文摘要:

      The assembly quality of aircraft large components often plays a decisive role in the overall assembly quality. With the development of augmented reality technology, the multi-source heterogeneous data involved in the assembly and docking process of large components can be visually displayed in graphical or digital form. This can assist process personnel in completing complex assembly tasks better. To overlay virtual information into the real world, accurate positioning of the poses of large components is required. However, in the actual process of wing-fuselage docking, the large components themselves and the assembly environment are relatively complex. This may lead to low positioning accuracy and computational efficiency in traditional tracking and registration methods. Therefore, a new tracking registration method has been proposed in this paper. Firstly, an image dataset is constructed based on Unity Perception. Then, feature point recognition and template matching are implemented based on Harris and scale invariant feature transform (SIFT). Finally, the matching threshold is optimized based on the fruit fly optimization algorithm. The effectiveness of the proposed method is verified through a wing-fuselage docking simulation experimental platform. The average total error percentage of the new method is only 2.006 9%. The method proposed in this paper solves the key problem of tracking, registering and positioning large components, and thus provides more reliable support for the further application of augmented reality technology in the field of aircraft assembly.